Runoff prediction and management models for the Njemps flats Baringo district Kenya

Abstract

'!his thesis examines the possibility of usirg artificially derived nmoff parameters to estimate the nmoff potential of natural rainfall on the Njenps Flats of Barirgo district (Kenya), am to obtain necessary hydrological data used to fonnulate guidelines for effective natural nmoff harvestirg am managementfor reforestation in the area. Simulated rainfall was used to derive artificial nmoff parameters which were subsequently applied to monthly rainfall data to obtain estimates of nmoff potential. Derived nmoff data were then used to detennine monthly nmoff probability, optinn.nnmicrocatchmentplot sizes, am someoptimal sil vicul tural practices, all of which were .found to closely agree with ~tions madeurrler natural rainfall conditions nearby. It was found that nmoff on the Flats is principally a function of stom characteristics (depth, intensity am frequency) which account for up to 90%of its variation. Runoff events can be predicted from the probability of effective stonns, Le., those exceedirg 7mrndepth am lOImtVhr intensity in the area. Rain can occur on the Flats in any month am with the low threshold rainfall of about 7mrn,the monthly nmoff probability ranges from about 0.2 in the driest monthto more than 0.77 in the wettest. Runoff harvesting can be a reliable reforestation support technique in the area. '!hough the meanannual rainfall of 640rnm on the Flats is not enoughto support tree growth, am a large I1r)isture deficit ran;JiIg from 960to 1600rnm per anrnnnexists in the area, it has the potential to produce 132mm of nmoff. '!his nmoff potential would require an optimum plot size ran;JiIg from 7 to 12 m2 to collect sufficient water to support the growth of seedliIgs through the dry season. SUcha plot size lmfortunately has the Potential to collect excess nmoff in the wettest 1OOt1twhshose potential effects such as erosion am seepage need further since the probability of runoff oc:x=urenceis higher from March to August, the fonner is the ideal plantiIg tiIne to ensure that the trees makeuse of all the wet season. However,the dry season between November am Febroa:rycieInanjg that a tree SPeCies like Prosopis with acceptable grcMtham Ilr)derate water use be used as its chances of smviviIg this drought are nnlchhigher than those of faster growiIg, high water-consuming species such as Eucalyptus camaldulensis am Ieucena leucocephala. '!he methodology of deriving optimumplot size and other infonnation needed for runoff managementfrom both artificial runoff parameters and natural rainfall as used in this study can be extended to other areas, provided the SPeCific variations in rainfall- nmoff relationships and other factors are fully established.