Laserfiche WebLink
<br />I <br /> <br />I <br /> <br />ABSTRACT <br /> <br />I <br /> <br />Individual flood hydrographs observed on <br />agricultural and range land throughout the United <br />States of America were described bya three-para- <br />meter mathematical model. The values of the three <br />hydrograph parameters were related to topography, <br />land use and rainfall associated with each flood event. <br /> <br />I <br /> <br />Three separate multiple regressions were <br />developed, each for predicting one of the hydrograph <br />parameters from three independent variables. <br />Thirty~six independent variables were tested including <br />many which have been suggested by previous workers <br />and which- are incorporated in current design methods. <br /> <br />I <br /> <br />I <br /> <br />Multiple linear regression, with three inde- <br /> <br />I <br /> <br />pendent variables, accounts for 80 percent and 61 <br />percent of the variance in the volume and peak rate <br />of runoff respectively. The third parameter a time <br />variable, which also defines the hydrograph recession, <br />can be predicted by a logarithmic relationship in- <br />volving only soils, land use and topography so that <br />about 60 percent of its variance is explained. <br /> <br />Due to the sample size being limited to forty- <br />seven flood events, the developed regressions do not <br />provide a rigorous formula for design. They are <br />presented rather to show that this approach prOvides <br />a simple method by which hydrographs could be <br />synthesized. The discussion suggests that a large- <br />scale study could provide the desired accuracy, and <br />indicates ways of achieving this. <br /> <br />ACKNOWLEDGEMENTS <br /> <br />I <br /> <br />The author wishes to acknowledge the guidance <br />of Dr. V. M. Yevdjevich in conducting this research <br />and his advice on "!he preparation of this dissertation. <br />Thanks are also due to the other members of the <br />author's research and dissertation committee for <br />reviewing the manuscript and for earlier discussions. <br />Dr. F. A. Graybill, Mr. W. S. Devenney and other <br />members of the Computing Service gave valuable help <br />in processing the data. <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />Thanks are due to the AgricultW'al Research <br />Service of the United States Department of Agricul- <br />ture for providing the funds with which the research <br />leading to the dissertation was carried out. Parti- <br />cular thanks is due to Mr. H. N. Holtan. Director <br />of the U. S. Hydrograph Laboratory of the Soil and <br />Water Conservation Research Division of the above <br />service. <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />I <br /> <br />The Department of Agricultural Technical <br />Services of the Republic of South Africa kindly afford- <br />ed the author leave to study in the United States of <br />America. The interest of Mr. E. D. Adler, Director <br />of its Natal Region. where the author will undertake <br />hydrological research, has been a great encourage- <br />ment. Thanks are also due to the innumerable per- <br />sonnel in bureaus of the United States for the courte- <br />ous and generous manner in which they provided <br />technical information and counsel. Special thanks is <br />due to Mr. H. O. Ogrosky, Chief of the Hydrology <br />Branch of the Engineering Division, of the U. S. Soil <br />Conservation Service, and to other engineers of the <br />Branch for valuable discussions and correspondence. <br /> <br />The author is most grateful to his wife and <br />family for the sacrifices they have made to enable <br />him to complete his studies. <br /> <br />ix <br />