To understand the research presented in this thesis a basic knowledge of ground penetrating radar is necessary. GPR is a geophysical tool used to image the subsurface. GPR systems consist of one transmitting antenna, one receiving antenna, a control unit, and cables connecting the antennae to the control unit (Figure 2). The cables carry a signal from the control unit to the transmitting antenna. The signal triggers an electromagnetic (EM) radio wave into the earth from the transmitting antenna. Portions of the EM wave are reflected back to the surface for a variety of reasons, such as contacts between stratigraphic layers or changes in the chemical or physical properties of soils or sediments such as water content or mineralogy (Jol and Bristow, 2003). The reflected energy is received at the surface by the receiving antenna and a signal is sent back to the control unit through the cables. The control unit converts the signal into information that is displayed on a computer as raw data (Conyers and Goodman, 1997; Wenell, 1998).

Figure 2. The transmitter antenna (1) sends EM waves into the ground and the receiver antenna (2) receives the EM waves that are reflected off underground layers. The information travels to the computer equipment by cables (3) and the computer equipment displays the raw data in real time on the screen (4).

The amount of time it takes the EM energy to travel from the transmitting antenna and be reflected back to the receiving antenna is referred to as two-way travel time. Two-way travel time is one factor used to determine the actual depth of the reflectors (the layers in the subsurface that reflect EM waves back to the surface). The other factor is the estimated velocity that EM waves travel through the soil. The estimated velocity can be determined by running a common mid-point (CMP) survey. CMP surveys are completed by separating the transmitter and receiver in evenly spaced increments and result in a profile plotting antennae separation, or distance, against two-way travel time (Figure 3). The profile can be used to calculate an estimate of the EM velocity in the shallow subsurface (Jol and Bristow, 2003). Depth is then calculated by dividing the travel time by the velocity of the EM waves.


Figure 3. Common mid-point profile from Cannon Beach. Velocity is calculated by dividing antennae separation (distance) by two-way travel time.

The frequency of the EM wave selected is a very important consideration in a GPR survey. Antennae frequency (measured in megahertz, or millions of cycles per second) determines the depth of penetration of the EM energy and the resolution of the GPR data received. GPR antennae frequencies range from 10 MHz to 1000 MHz (Jol and Bristow, 2003). High frequencies have shorter EM wavelengths which result in shallower depths of penetration into the ground meaning deeper objects and layers of interest will not be detected. However, shorter EM wavelengths are able to detect smaller objects and changes in sedimentology in the ground and result in higher resolution GPR data. Thus, there is a decrease in depth of penetration with the increase in resolution.

Step interval, another important consideration in a GPR survey, determines the distance the antennae are moved between each trace. Step interval affects the horizontal resolution of the GPR data. GPR surveys using 100 MHz antennae should have a step interval of around 0.25 m (Jol and Bristow, 2003). If the step interval is too large, the transect will not provide adequate information to identify the reflections.

To collect GPR data, the antennae are moved along survey lines. EM waves are sent into the ground at evenly spaced locations along the survey lines. The receiving antenna receives reflected EM waves at the ground surface and the data is sent to the control unit. As the energy is transmitted into the ground at intervals along the survey line the reflection traces are arranged side by side to form a reflection profile of the survey line. The darkly colored areas on a reflection profile are layers that are strong reflectors. These reflectors may indicate the presence of a subsurface contact or buried object. Two-way travel time and depth are shown on the y-axis of a reflection profile and the surface position is indicated on the x-axis.

GPR data is stored digitally on a computer in real time. This means the reflection profile appears on the computer screen as the data arrives from the receiving antenna. Hence, GPR data can be interpreted in the field (or in the lab after the survey is completed). GPR data must sometimes be processed before it can be properly interpreted and two-way travel time and velocity must be converted to depth estimates. There are many GPR processing programs and techniques available. Typical data processing includes some type of gain and filtering. Gain functions compensate for the loss of signal strength with increased depth. A common type is automatic gain control (AGC) which equalizes the amplitudes all the way down each trace while spherical and exponential gain compensation (SEC) applies a linearly increasing time gain combined with an exponential increase in the gain (Jol and Bristow, 2003). Filtering is used to remove or emphasize different types of spatial variations (Jol and Bristow, 2003). Topographic corrections should be applied as changes in relief will affect the apparent depth and appearance of reflections.


Sensors and Software pulseEKKO 100 and 1000 GPR systems were used to collect the GPR data for this study. The equipment configurations, survey and data processing parameters for each transect are shown in Table 1. The pulseEKKO 1000 was equipped with an odometer wheel to accurately measure distance and for rapid, continuous data collection. Using the pulseEKKO 100, the antennae were moved in one-meter step intervals. GPR lines, and step interval while using the pulseEKKO 100, were measured to the nearest one cm on 100 m tapes.

Elevation data was collected along three transects in Cannon Beach using a laser level with three cm accuracy over 100 m distances (Figure 4). Coring was completed by Dr. Curt Peterson of Portland State University at each site in Seaside with a gouge core auger (2.5 cm diameter) to ground truth the GPR data (Figure 5).

Figure 4. A laser level was used to collect elevation data in Cannon Beach.   Figure 5. Core samples were collected at each site in Seaside.


Research was conducted in the low-lying coastal communities of Cannon Beach and Seaside, Oregon. Study sites were selected to investigate the extent of paleotsunami inundation and the magnitude of wave run-up.

GPR data was collected at four sites in Cannon Beach. Data was collected on roads with across-barrier (Second Street) or along-barrier (Larch and Spruce Streets) orientations and in a back-barrier wetland (Spruce Street marsh) to establish barrier stability and to determine the direction of overland paleotsunami flow (Figure 6). Elevation data was collected for Larch, Second, and Spruce Streets.

cannon map

Figure 6. The four sites in Cannon Beach were located on roads and in a back-barrier wetland.

Four sites were selected in Seaside as well (Figure 7): 1) the north end of Pine Street (HORN), 2) north of the 12th Avenue bridge along the Necanicum River (12TH) , 3) east of North Roosevelt Drive approximately 150 m north of 17th Avenue (east of Seaside High School and M & F Plumbing LLC) (PLUM), and 4) east of Avenue K (KAVE).

seaside map

Figure 7. The sites in Seaside were located on the shores of river channels and at the site of an assumed paleotsunami pour-over fan.

At the Avenue K site, data was collected parallel and perpendicular to an assumed paleotsunami pour-over fan. Data was collected on the shores of river channels where there was observed or assumed tsunami flow during the 1964 tsunami at the other Seaside sites. Two transects were run at each site, perpendicular to one another to gather data from each site. Core samples were obtained at each site in Seaside at the beginning and end points of each transect and in the middle of each where the two transects crossed.


The GPR data was processed using Sensors and Software pulseEKKO software. Dewow and automatic gain control (AGC) were used to process all of the profiles. Topographic correction was applied to the profiles for which transect elevation data was collected during the GPR survey. The data is presented as wiggle trace plots.

[Previous Section]