by Charles French (University of Cambridge)
The Marakwet field workshop in December 2013 enabled a geoarchaeological assessment of the landscape character of this region and some initial soil sampling. Particular attention was paid to three areas: first the farm compounds, fields and slope off-take irrigation system occupying the lower slopes to the northwest of the village of Tot; second the River Embobut valley at the base of the slope over a distance of about 4km eastwards; and third the wider valley floor to the southeast-east of Tot and the main River Kerio floodplain and low terrace area towards Pokot where the Canadian Red Cross has extensively cleared land for a new irrigation system and multiple field plots over an area of some 250 hectares.
A more detailed report of these observations and the soil profiles described has been placed on file and will be incorporated into future research. The following provides a brief summary of the observations so far:
The lower slopes north of Tot
The River Embobut descends rapidly from the c. 2,600 metres above sea level escarpment to the north. From about the 1,100-1,000 metre level there is an extensive series of irrigation ditches or ‘furrows’ defined by stone and concrete or earth walls that follow hillside contours down-slope to feed a variety of house/farm platforms and associated small plot fields. These have been mapped, recorded and discussed in various previous publications and especially in the most recent phases of work conducted by Moore and Davies. The water flows year-round as it is fed by the rain- and spring-fed system on the escarpment above. The furrows must require constant repairs because of collapse and erosion factors on the relatively steep slopes (c. 20-45 degrees), as well as regular cleaning out of the substantial and rapidly accumulating fine sand deposits in every furrow.
The new Canadian Red Cross irrigation system has built a new concrete and stone spillway structure to take a proportion of the water from the River Embobut at the c. 1,100 metre level via a 30cm diameter pipe. Significantly after one rainy season, the spillway was already about half-full of re-deposited fine-medium sand material. It will certainly require regular flushing/cleaning out.
The lower slope zone at about 800-600 metres is characterised by thick, gently undulating, lobate zones of red fine sandy-silt colluvial material. These colluvial fan areas are extensively cropped. One east-west erosion gully recorded as ‘Lagam’ has revealed significant incision of over 3m that has occurred since 2007 (Nelson, pers. comm.). This gully cut revealed a probable thin (c. 12-15cm) buried soil at its base that was composed of brown fine sandy silt loam developed on highly weathered mica-rich basaltic hard-rock. Above this were three thick units of hillwash accumulation: reddish brown, gravel free, fine sandy silt; reddish brown medium-coarse sand with abundant fine gravel; and an increasingly stoney/gravelly but similar reddish brown medium-coarse sand; all below about 90cm of modern, brown, fine sandy silt topsoil and large stone cobbles of the present day field system. The main horizons of this profile were spot sampled for micromorphological and geo-chemical analyses.
The Embobut river valley and floodplain
From the base of the slope by the road bridge for a distance of c. 4km southwards the River Embobut and its available river-bed profiles were fieldwalked and six profiles were recorded and two were spot sampled for micromorphological and geo-chemical analyses (RP01 and RP06).
Profile RP01 was located about 300m southeast of the road bridge beyond the foot of slope. Here there was a large embayment of over 4.25m in depth which began to become incised from c. 1978 (Nelson, pers. comm.). This profile revealed a basal c. 50cm thick alluvial soil beneath a c. 75cm accumulation of alluvially derived organic sandy silt to fine gravel deposits, with almost 3m of hillwash derived material above composed of mottled brown to reddish brown fine sandy silt with abundant fine pebbles.
Within about half a kilometre to the southeast from the foot of the hill-slopes this colluvium over alluvium profile ceased to be evident and was replaced by profiles that were consistently composed of riverbed coarse sands/cobbles and fine sandy to silty clay alluvial materials. In addition, the c. 3-4m of incision was now stepped in most places to reveal two terraces, the lower one at about 1.5-2m above the present day river-bed, and the second, c. 1.5-2m higher terrace on the outer edges of the river valley, extending to the undulating hard rock geology some 100-200 metres distant to the east and west. These terraces are regularly under (cash) crops today; although there is the risk of loosing the crops to flash flooding.
The lower river terrace deposits are consistently composed of greyish to yellowish brown, fine sandy/silt loam and fine gravels which reach a thick (c. 40-50cm) stabilisation zone at a depth of c. 1/1.5m below the present day highest valley ground surface (or top of the second, later terrace). This upper surface of the first terrace is a dark greyish brown, organic sandy loam with some columnar to blocky structure evident, and has the appearance of a considerable period of stability with minimal or no seasonal alluvial aggradation.
Alluvial deposition again resumes to form the make-up of the second/higher terrace. This is comprised of multiple alternating horizons of mottled orange silt, pale yellowish white medium to coarse sand, fine sandy silts and bedded fine gravels over a thickness of c. 1-1.5m.
This two terrace alluvial and river-bed sequence was sampled near Mr Atomic’s first valley farm at profile RP06. As evident at profile RP01, there is a c. 50cm thick, well structured, silty clay alluvial soil at the base of the sequence, which contains much amorphous iron mottling. The latter is suggestive of much in situ standing vegetation, possibly indicative of a former wet, marshy alluvial soil in the valley bottom.
The River Kerio floodplain
The present day muddy River Kerio is well incised into the surrounding floodplain to a depth of c. 3.5-4m through reddish brown fine sandy silt alluvial material. River-edge profiles were too overgrown to record properly, at least at this time, but the alluvial topsoils were predominantly fine-medium sandy loams. There is an extensive floodplain zone of at least 500-1000m to either side, which is the subject of the Red Cross irrigation scheme. At least 250 hectares has just been cleared of acacia scrub and the irrigation pipes have been laid every c. 100m along the long axis of the floodplain.
Soils of the Tot region
Field observations suggest that there are two main types of soils present: ferrallitic and sandy alluvial floodplain soils. Primarily, the soils of the Tot area on the footslope, lower slopes and parts of the Embobut floodplain are ferrallitic soils. These are thick, freely draining, weakly acidic soils dominated by iron and aluminium sesquioxides with quartz sand and kaolinite clays, with the silica removed from the clay, in an oxic B horizon (Bridges 1978, 82-3). These soils are characterised by their strong red colours and cementation with amorphous iron oxides. Depending on the degree of vegetation cover and leaching (related to rainfall and humidity), these soils typically have little or no reserve of weatherable minerals and consequently usually exhibit low fertility. This latter point is a potentially significant one with respect to crop choice, growth and production without irrigation and nutrient replenishment by fertilisation.
A possible model of landscape development to test
From the initial geoarchaeological evaluation just carried out, it is possible to suggest a sequence of soil/sedimentary history for the Tot-Embobut-Kerio-Pokot landscape over the last 10,000 years or so. Obviously this will need further detailed mapping, description, sampling and analyses, and importantly dating using radiocarbon and OSL methodologies, and comparison with the known ethnographic and historical narratives, but is sufficient to present a model of landscape development to test.
With the advent of the Holocene and post-glacial warming, afforestation took place over the whole landscape with concomitant ferrallitic soil development under a rain-fed regime. Some impact on the woodland followed, perhaps for some thousands of years, through dispersed and light-touch human activities, but was sufficient to start some colluvial processes and eroded soil accumulation in the toeslope areas and alluvial aggradation in the Embobut and Kerio River valleys and floodplains, up to a thickness of about 50cm. This suggests that there was some tree clearance and soil disturbance on the lower slopes northwest of Tot which led to some erosion and downslope movement of ferrallitic soils via rainfall/stream action and overland flow. When these minimal, slow and gradual erosion/aggradation processes began is open to debate at this stage given the largely unknown record of human activity prior to c. 300 years ago in this region. Then there was a major change from a minimally erosive system to one of much greater severity. This involved greater depths (c. 1-1.5m) and coarser soil erosion on both the footslope and alluvial floodplain areas. There was then a relatively lengthy phase of stabilisation, probably representing several hundred years, represented by the lower terrace floodplain soil in the Embobut river valley. Subsequently there was an abrupt change which appears to have involved major clearance and disruption of the hill-slopes leading to up to 2m of hillwash at the base of the slopes and 1.25-1.5m of coarse alluvial accumulation in the river valley floodplain. This may be associated with the intensive development of hillside farms and the furrow irrigation system, but this remains to be proven. All of this eroded material took on a coarser character suggesting greater intervention on the slopes above and exposure of some weathered bedrock material, and the Embobut river began to change coarse more rapidly and regularly, exhibit very different flow velocities and volumes of water as seen in the alternating fine to very coarse matrices accumulating, and build sand/gravel levees. This material forms the upper terrace of the Embobut river today.
Up until now the system has been one of erosion and accumulation, but it now changes to one of incision and removal in the lower slopes and Embobut river floodplain. Local farmers (Mr Atomic, Nelson and William, pers. comm.) indicate that this process began very recently, from the mid-1970s. In places here has been over 4m of incision and new gullying in both the toeslope and in the floodplain zones of the landscape. New channels regularly form and re-form in the Embobut floodplain, and both first and second terrace deposits are often scoured away by large volumes of floodwaters, ruining both new fields and cash crops on the floodplain margins, with coarse sand/fine pebbles levee creation in the channel floor.
The most recent event is the construction of the Canadian Red Cross irrigation system last year to take water from the Embobut river upslope to a large area of the Kerio floodplain southeast of Tot to enable crop agriculture on a new extensive and large scale not seen before in this landscape. Although this system is not yet operational, I understand that it will be maintained and fertilisers provided to local farmers for a period of only three years. What will happen afterwards in terms of maintenance, sustainability and success of the project will bear careful observation and analysis.
Things to do
There are a number of unknowns that require further fieldwork and analysis to understand both the long-term landscape history of this area, as well as how this system has functioned so well over the past c. 100-300 years.
The spatial extent and depths of colluvium and alluvium should be recorded and mapped with respect to current land use and past/modern settlements. The mechanisms and derivation of these deposits need to be understood, and relative chronologies of deposition determined, and related to the known ethnographic narratives.
The nutrient base and elemental components of the modern soil system and colluvial/alluvial sediments and any buried/modern soils present need to be established as an index of fertility and ability to grow/sustain particular food crops. On the face of it, this landscape system is successful because of irrigation and both small scale and shifting agricultural practices, not because the ferallitic soils present are especially fertile in their own right.
A hydrological model of the current river and irrigation furrow system should be established. The water quality and nutrient status of the River Embobut and the irrigation furrow system, both on the slopes around Tot and in the valley below, need to be established. This may already be available from the current Canadian Red Cross irrigation project.
It is crucial to assess the impact and potential sustainability of the Canadian Red Cross water harvesting scheme on the furrow system and its associated fields, well as the large area of new agricultural plots on newly cleared land in the River Kerio floodplain, now and over the next three to five year period.
Finally, as the existing farming settlement system around Tot and Pokot has been so well studied over the past 35 years from an ethnographic narrative perspective, there is an unrivalled opportunity to use soil and ground-truth micromorphological and geo-chemical analyses to investigate the detailed use of space and activities represented in the farmsteads.
I would like to thank our hosts and guides in the field at Tot, especially Hellena, Nelson, William, Timothy Kipkeu and Mr ‘Atomic’, as well as all the staff and ‘Mama’ of Prof Henrietta Moore’s expedition house, and the African World Vision centre in Tot. I would also like to thank the British Academy International Partnerships scheme for funding the project.
Bridges, E.M. 1978. World Soils (2nd edition). Cambridge: Cambridge University Press