Document Type


Date of Degree

Spring 2014

Degree Name

MS (Master of Science)

Degree In

Civil and Environmental Engineering

First Advisor

Witold F. Krajewski


After the flood event in 2008, Iowa Flood Center (IFC) designed a new water level measuring system, Ultrasonic Stream Bridge Sensors (USBS), to monitor stream water level in Iowa. The system is composed of an ultrasonic, a GSM cell modem, solar panel, and battery, and an internal temperature sensor; all components assembled in a weather proof box. The USBS are designed to be mounted in a bridge crossing and uses speed of sound to sense the distance from the sensor to the water surface. USBS are inexpensive compared to other system of water level measuring systems. However, ultrasonic sensor in USBS is very sensitive to variations in air temperature and changes in air density between the sensor and the water surface, which can be a major source of error in distance measurement. To reduce the effect of change in air density on the distance reading USBS internally compensates distance using temperature measured by its internal temperature sensor. IFC specifies that the sensor measures to an accuracy of 1% of the measurement range. However, more than three years of water level data collected by the sensors shows that there were fictional water level fluctuations to the order of +/- 7cm on average in most of the sensors. Spectral analysis done on the data also showed that the fluctuations have a strong diurnal cycle behavior. The cycle was stronger on South facing sensors. To reduce the error two methods of compensation were developed based on previous literatures and Senix Corporation advice. The first compensation method undoes the internal compensation and compensates the error using air temperature from nearby weather station. This method was applied to USBS who have close air temperature measurement from weather stations and to an experimental USBS installed in Iowa City airport. The method reduced the fluctuations by an average of 2cm in most of the sensor. The second method predicts local air temperature of the stream channel based on energy balance of the channel. The model predicted channel air temperature slightly less than air temperature from nearby weather station and we are able to reduce the fictional fluctuations on the order of 1cm.


x, 61 pages


Includes bibliographical references (pages 60-61).


Copyright 2014 Tesfalem Tekle