Soil erosion, iron smelting and human settlement in the Haubi Basin, north-central Tanzania. (Special section).

Lane, Paul ; Mapunda, Bertram B.B. ; Eriksson, Mats 等

The Haubi Basin, situated in the Irangi Hills of Kondoa District, Dodoma Region, Tanzania (FIGURE 1), exhibits some of the most extreme examples of erosion and associated sedimentation in the region (Christiansson 1981; Payton et al. 1992) (FIGURE 2). The severity of the problem, with its consequent loss of productive land, has been known since the beginning of the 20th century (Kannenberg 1900; Obst 1915). Recognition of the extent and severity of the issues encouraged the British colonial government to introduce a variety of measures aimed at soil conservation in the 1930s (Fosbrooke 1950). These included the construction of check-dams, contour bunds and contour planting with sisal. After independence, a second programme of soil conservation measures was launched. These were initially similar to those employed under colonialism, but later included de-stocking, resulting in the removal of some 90,000 livestock from the most severely eroded areas. As in the colonial period, recent overexploitation of land resources by local populations was regarded as the primary explanation for the continuing soil erosion.

[FIGURES 1-2 OMITTED]

Recent investigations involving a combination of geomorphological studies with OSL dating of sediments, however, have identified two main phases of erosion, both of which commenced well before the mid 19th century, and are therefore unrelated to recent land use practices (Eriksson 1998). The first occurred between 14,500 and 11,400 BP, coinciding with the end of the Pleistocene. Proxy environmental data from throughout the region indicate that this period was significantly wetter than today. The second phase appears to have commenced around 900 years ago, with a new phase of gullying being initiated sometime after AD 1400.

Since farming and herding were well established across the region by 2000-1800 BP, an increase in human settlement, iron smelting and/ or livestock grazing could have contributed to the initiation of the second major phase of soil erosion some 900 years ago. In an effort to assess these hypotheses, a programme of survey and test-excavations was begun in 1999, aimed at collecting data on the dating, distribution and topographical locations of sites of different periods, and recovery of samples that would allow the reconstruction of subsistence strategies and metal-producing technologies.

Three seasons of archaeological fieldwork have been completed. A fourth is scheduled for 2002. Over 50 separate artefact scatters have been located, which range in date from the ESA to the 19th century. Three rock-shelters containing archaeological deposits, two with rock paintings, have also been found. Over 70% of the scatters can be attributed to the Iron Age (i.e. post 2000 BP). These comprise surface spreads of pottery, iron slag and tuyeres, with occasional pieces of house daub. Most occur on the middle pediment slopes above Lake Haubi; have been cut by gullying up to 30 m deep; and are severely deflated. The only securely dated Early Iron Age site (Haubi 16) occurs at the head of the gully system on the upper pediment slopes, whereas the bulk of the Later Iron Age (c. 1000-200 BP) sites are found lower down. Sites dated after AD 1800 occur on both the middle slopes and lake basin floor. A suite of radiocarbon assays on charcoal from in situ deposits or lumps of iron slag recovered from the Haubi and adjacent Mwisanga basins, when calibrated using the OxCal programme, tends to support the impression gained from the survey data that there was a marked increase in settlement and smelting activity after c. AD 1300.

Whether this apparent upsurge in activity was responsible for triggering soil erosion remains uncertain, however. While there was certainly an increase in iron smelting, metallurgical analysis of smelting debris suggests that the technology was very fuel efficient. In addition, former local iron-smelters claim to have used only three to four hardwoods for charcoal (FIGURE 3). As these species make up only a fraction of local forests, iron smelting by the late 19th century probably had only minimal impact on the environment. Establishing the antiquity of such `species selectivity' will be a focus of the next phase of research. Equally important will be to account for the increase in smelting activity from c. 1500 AD, and whether this was associated with an intensification of regional trade. One possibility, as yet unverified but suggested by similarities in pottery, is that the iron smelters of the Irangi Hills were part of a network which linked them with grain-producing areas over 100 km to the north around Engaruka, where extensive fossil fields and irrigation systems dating mostly to the 15th-17th centuries still survive (Sutton 1998).

[FIGURE 3 OMITTED]

Acknowledgements. We are grateful to the Tanzania Antiquities Unit and COSTECH for permission to undertake the archaeological research, which has been funded by the British Institute in Eastern Africa. We would like to thank all members of the MALISATA programme who have freely shared their data with us, especially Dr Robert Payton and Professor E.K. Shishira, along with Stephan Woodborne and the Quaternary Dating Research Unit, Pretoria. The metallurgical analyses were conducted by Bertram Mapunda at the Institute of Archaeology, London, while on a Commonwealth East African Visiting Scholarship organised by the School of Oriental & African Studies. We are especially grateful to the Institute's director, Professor Peter Ucko for having helped facilitate this. We would also like to thank all of the British, Tanzanian and Swedish students who have assisted with this work, and the residents of Haubi for being so accommodating.

References

CHRISTIANSSON, C. 1981. Soil erosion and sedimentation in semi-arid Tanzania: studies of environmental change and ecological imbalance. Uppsala: Scandinavian Institute of African Studies and Department of Physical Geography, University of Stockholm.

ERIKSSON, M.G. 1998. Landscape and soil erosion history in Central Tanzania: a study based on lacustrine, colluvial and alluvial deposits. Stockholm: Department of Physical Geography, University of Stockholm. Dissertation series 12.

FOSBROOKE, H. 1950. The fight to rescue a District, East African Annual for 1950: 61.

KANNENBERG, H. 1900 Reise durch die Hamitischen Sprachgebiete um Kondoa, Mitteilungen von Forschungsreisenden und Gelehrten aus den Deutschen Schutzgebeiten 13: 144-72.

OBST, E. 1915. Das abfluBlose Rumpfscollenland im nordostlischen Deutch-Ostafrika. Bericht fiber eine im Auftrag der Hanburgischen Gegraphischen Gesellschaft in der Jahren 1991/12 usgefuhrte Forschungsreise, Mitteilungen der Geographischen Gesellschaft in Hamburg, Band 29.

PAYTON, R.W., C. CHRISTIANSSON, E.K. SHISHIRA, P. YANDA & M.G. ERIKSSON. 1992. Landform, soils and erosion in the north-eastern Irangi Hills, Kondoa, Tanzania, Geografiska Annaler 74A: 65-79.

SUTTON, J.E.G. 1998. Engaruka: an irrigation agricultural community in the northern Tanzanian Rift Valley before the Maasai era, Azania 33: 1-37.

Mats Eriksson, Lane, British Institute in Eastern Africa, PO Box 30710, Nairobi, Kenya. pjlane@insightkenya.com Mapunda, Archaeology Unit, University of Dar es Salaam, PO Box 35050, Dar es Salaam, Tanzania. bbbmapu@udsm.ac.tz Eriksson, Dept for Natural Resources and the Environment, SIDA, 10525 Stockholm, Sweden. mats.eriksson@sida.se