Hydrological modelling in the meso scale semiarid region of Wadi Kafrein / Jordan -The use of innovative techniques under data scarcity

Abstract

In order to effectively plan and manage the water resources of a country or geographical region, available water resources must be quantitatively assessed and hydrological processes must be known. Although surface water is ordinarily considered a primary source of water, the Middle Eastern focus has been primarily on groundwater. Attention to surface water, runoff generation processes, and overall catchment modelling has typically been paid little attention. However due to hydrological conditions and population increase, countries like Jordan are considering the entire spectrum of water resources and the quantitative assessment and characterization of hydrological resources are becoming ever more important. Hydrological modelling in the arid to semi arid catchment of Wadi Kafrein (161 km²) was the objective of this study, and in order to complete this objective, detailed hydrological investigations were performed there. The catchment is characterized by a wide range of climatic differences, topographic variations, and spatial land uses. A physically based, spatially distributed, hydrological model was prepared within the framework of this dissertation. The aim of this dissertation is a detailed quantification of the water balance of the study area with emphasis on (1) runoff generation mechanisms and (2) the resulting transmission losses. A secondary aim of this study was for the model to assist local decision makers in solving water resources management problems. Due to the large variability in hydrological parameters of the catchment area, the model was intended to be physically based and spatially distributed. The type of physically based model which was selected for the Wadi Kafrein catchment is the TRAIN-ZIN model. In order to meet the requirements of this model, data which was of high spatial and temporal resolution was obtained and a comprehensive hydrological database was prepared. The geometry of the catchment (Digital Elevation Model, DEM) was prepared using Cartosat-1 satellite images with a resolution of 5 m. The spatial variations in land use and soils were graphically shown in respective land use and soil maps by using multi-temporal ASTER satellite images. In order to calibrate the model, runoff measurements were required, and therefore, the topography of the surface water reservoir behind the Kafrein dam was surveyed during a dry period using high resolution differential GPS measurements. From the resulting elevation model, a rating curve for the surface water reservoir was prepared. This allowed the quantification of surface runoff by water height determination. The water height in the reservoir was measured continuously using data loggers. The monitoring period extended from November, 2007 until December, 2009. Several sub-wadis with catchment areas of 0.3 km2 to 7 km2 were instrumented for high temporal rainfall and runoff measurements. The numerical model was parameterized, calibrated, and validated using the measured data. The model was calibrated and validated using the Differential Split Sample Test approach. The water components of the two consecutive hydrological years were quantified and spatial distribution maps were prepared for every water component on an event basis. The results of a sensitivity analysis indicate a strong effect of soil depth and soil infiltration rates on the generated runoff amounts while transmission losses are mainly affected by channel length, channel width, and the depth to active alluvium. Runoff generation mechanisms were quantified in detail for the monitored runoff events. The results show that Infiltration Excess Overland Flow (IEOF) is the dominant runoff mechanism in the study area, which is also known to be the dominant mechanism in other arid and semi arid regions. On the other hand, and despite that fewer events were generated due to Saturation Excess Overland Flow (SEOF), the results show that, quantitatively, runoff is mainly generated due to SEOF. For similar amounts of rainfall measured in two different storm events, the volume of the generated runoff with SEOF as the dominant runoff generation mechanism was ten times more than an earlier event with a similar measured rainfall amount. This earlier event was characterized by IEOF. This observation may be attributed to the rainfall intensity, the antecedent soil moisture, and the lag time between the storm events. Transmission losses were also quantified on an event basis and on an annual basis. Transmission losses ranged from 18-44% of the generated runoff on an event basis while the average transmission losses were 24% and 26% of the generated runoff in 2007/2008 and in 2008/2009, respectively. The maximum runoff coefficient was 4% in 2007/2008 and 11% in 2008/2009. Recharge was higher when SEOF was the dominant mechanism and lower when IEOF was the dominant mechanism. In order to attempt to predict the impact of climatic patterns variations and the impact of urban expansion and land use changes on the water balance components of Wadi Kafrein, a new approach has also developed within the bounds of this dissertation. This new approach takes into consideration precipitation characteristics and temperature increases, including the wettest and driest years on the available records. In total, 24 climatic scenarios were developed and the results were further validated by applying continuous modelling for the hydrological years from 2002 until 2007. The results of climatic scenarios show that runoff coefficients range from 4% in very dry years to 21% in very wet years. Furthermore, an increase in temperature of 1-3 °C will slightly decrease recharge and runoff. Urbanization expansion in Wadi Kafrein will mainly increase the volume of generated runoff and decrease the recharged water. As a further result of the research described herein, indications are that previous estimations of runoff and recharge in the Wadi Kafrein were too low and evapotranspiration was too high. The results from this study indicate that on an average year, runoff is approximately 6.4 MCM and recharge is around 21 MCM. Recharge equations were also developed to estimate recharge based on annual rainfall.