Library Header Image
LSE Theses Online London School of Economics web site

A distributed simulation model of a reconstructed ancient water harvesting system in the Negev Desert.

Lee, Michael David (1990) A distributed simulation model of a reconstructed ancient water harvesting system in the Negev Desert. PhD thesis, London School of Economics and Political Science.

Download (218MB) | Preview


This thesis addresses the problems of evaluating a water harvesting catchment system in the arid environment of the Negev Desert, Israel. A general interpretation of arid hillslope hydrological response is developed called Contiguous Area Contribution. Whilst agreeing with humid area concepts of partial area contribution, it focusses on the different nature of arid slope contributions to stormflow where frequent and rapid overland flow generation intercepts channels by downslope expansion and extension of flow-lines. The physical system at Avdat is geometrically represented as a flow net of hillslope and channel cascades for kinematic routing. A survey of the micro-morphology and surface materials enables hillslope areas to be classified into six broad units on the basis of their textural characteristics. These units are used to quasi-distribute process data sampled at locations within them. The process of infiltration is discussed and different mathematical models examined using results from a limited number of field measurements with runon/runoff apparatus. The best-fit is provided by the storage model of Green and Ampt and not the Hortonian models of Philip and Kostiakov. Initial infiltration rates vary from 85 to 18 mm hr-1 and steady- state rates from 60 to 6 mm hr-1. The inclusion of detention storage in early time period observations is shown to be a significant problem in modelling. For flow routing, an alternative approach to assuming sheetflow is presented using cross-slope microtopographic profiles to provide an estimate of surface geometry for flow across micro-rough surfaces. Measurements of flow velocity for different discharges are made using the runon/runoff apparatus. By assuming sheetflow, a lower value of Manning's n is predicted if velocity and flow dimensions are known. For a given n, the assumption of sheetflow predicts a lower velocity for a given discharge. The values of n derived at Avdat range from 0.18 to 1.36 with a mean of 0.64 when a micro-rough surface is retained, and 0.12 to 0.62 with a mean of 0.36 if sheetflow is assumed, values considerably higher than those usually adopted from channel studies. A distributed model is developed to handle surface runoff processes at a range of scales from the small plot to the complete catchment. In a detailed sensitivity analysis, the range of physical and process parameters derived for Avdat show sensitivity of the runoff processes to particular parameters and their combination. For the selected range, flow boundary shape is the most significant influence on the shape of flow hydrographs resulting in quicker, higher peaks if a micro-rough flow geometry is assumed. The model is used at the plot scale to examine the problem of including detention storage in infiltration model parameters, at the cascade scale to show the effect of runoff production enhancing and inhibiting slope areas, and at the sub-catchment scale to assess the predictive ability of the model using the limited process parameter data-set. With three sub-catchments, prediction errors were good volumetrically ranging from 6% to 14% for the high intensity, short duration rainstorm, but deteriorated for the long duration, varied intensity storm with predicting high overestimates. Three sub-catchments consistently under-predicted and one over-predicted. In most, the rising and falling limbs were lagged relative to the observed hydrographs.

Item Type: Thesis (PhD)
Uncontrolled Keywords: Hydrology, Water Resource Management
Sets: Collections > ProQuest Etheses

Actions (login required)

Record administration - authorised staff only Record administration - authorised staff only


Downloads per month over past year

View more statistics