Gordan Buckingham – Investigating Near-Shore Submarine Groundwater Discharge (SGD) in the Ka’a’awa Valley, HI

Gordan Buckingham
Appalachian State University

Submarine Groundwater Discharge (SGD) is freshwater that flows underground and then surfaces in the submarine environment. It is a well-known, but not much studied phenomenon.4  It is important for several reasons, including managing water resources and understanding the negative effects of pollution on coastal zones. SGD is a pathway for material transport that can have a negative impact on coral health and the overall marine environment.1

SGD is an observed phenomenon in Hawaii.1,4 One major study of SGD was conducted in Kahana Bay, just north of the Ka’a’awa Valley, using seepage meters, geochemical tracers, and salinity meters.2  Aerial thermal infrared imagery has been used in Hawaii to identify diffuse and point source SGD plumes off the coast.3

The study area is the near-shore environment off the Ka’a’awa Valley watershed. The shore in the watershed is about 1.6 kilometers long. It is apparent anecdotally that the precipitation input in the valley does not equal the output from the streams. One possible explanation is that the precipitation is percolating into the groundwater and resurfacing in the ocean.  The purpose of this project is to identify possible areas of SGD using various methods, as well as identifying stream outlets where a freshwater influence is present. This project was conducted June 18th – 27th.



A variety of tools were used to attempt to locate and characterize SGD. The purpose was to corroborate any findings through multiple methods. These methods include utilizing Ground Penetrating Radar (GPR), conductivity and temperature meters, and thermal infrared imagery. Each is discussed separately.

(1)    The instrument used to measure conductivity and temperature in the water was the Solinst Levelogger 3001. Conductivity can be set to be taken in microsiemens (μS) or millisiemens (mS). Temperature is taken in Celsius. Conductivity measurements have been used to identify SGD in many environments.5

For off-shore data collection, two Leveloggers were attached to the bottom of two kayaks. The Leveloggers were synced up with the GPS devices to collect data at 10 second intervals.  Paths were taken along the shoreline at varying distances from the shore. Data was collected over three days (June 19, 21, and 24). One of the Leveloggers collected lower conductivity values for the same area than the other Levelogger, so the two needed to be calibrated. Salt water dilutions were used, along with a more accurate handheld conductivity meter, to correct the two Leveloggers. Even with calibration, the two Leveloggers produced different values for the same area. This could be attributed to the choppy surf and to the different time of day the values were taken.

One Levelogger was placed ~10 yards upstream from the high tide line where the main perenniel stream lets out from the valley. The purpose was to see whether there is a tidal influence to make the water brackish.


Instrument used to measure temperature and conductivity.


A handheld conductivity meter was used for samples on the beach for pore-water and in the marsh behind the road. The instrument used was an Oakton PC 650. It was also used as a control to calibrate the two Solinst Leveloggers. At many points, streams hit the sand and percolate down into it. Pore-water samples were taken to see whether the water disperses outward in the sand or whether it stays relatively channelized until it runs into the ocean. Pore-water samples were also taken at points not close to any stream channel to see whether the freshwater from the swamp areas across the road percolated under the road.

All but one stream percolates into the sand once they flow to the beach.


(2)     A handheld Garmin eTrex Venture HC was used to collect GPS waypoints and tracks. Garmin eTrex generally have an accuracy of + or – 4m.  The DNRGPS Open Source utility from the Minnesota Department of Natural Resources was used to import the Waypoints and Tracks. Points were then plotted in Arcmap 10.1.

(3)  The Ground Penetrating Radar (GPR) used was Geophysical Survey Systems, Inc.’s (GSSI) SIR-3000 with the 400 and 200 MHz antenna. The GPR was used along the roadside by the shore. The water table was visible, along with some dips where concrete culverts crossed the road. The purpose was to see whether the water table rises with any possible incoming groundwater flowing underground toward. A rise in the water table would suggest more moisture in the soil, possibly leading to a seep. The water table appeared consistent along the entire study area. Limitations to this could be the antenna used. Other wavelengths may pick up other features more distinctly.

Example of how a GPR reading appears.

GSSI SIR-3000 Ground Penetrating Radar


(4)   The thermal camera used is a FLIR Tau 640 attached to a DJI Phantom Quadcopter.  The camera is 640×512 VGA format with a pixel size of 17 microns. It is set up to be motion activated and has no timestamp, so pictures are taken at a variable rate. The camera also lacks GPS capabilities. Due to this and the lack of identifiable features in the study area, georeferencing and mosaicing the pictures to get a larger view of the area is difficult, if not impossible. If defined flight paths and more ground reference points were used, the images could be georeferenced to obtain a thermal image of the entire study area. ArcMap 10.1 was used to stretch values through the ‘interactive histogram stretch’ interface.

DJI Phantom Quadcopter





(1) No scale can be inferred from the thermal images, so the images are unusable unless parts of the shore are in the field of view for reference.

The red represents more emitted heat. For scale, the two figures in the top right are people.

Change in color gradient is likely waves crashing onto the shore.


Raw image shows perennial stream channel flowing into the ocean and mixing with the saltwater.

Contrast adjusted image.



         (a)  The conductivity data collected off-shore with the kayaks and from wading may be heavily influenced by the rough surf mixing the water in much of the study site. The time of day the data was collected may also give different readings for points collected in the same area. The two units we used were miscalibrated, so that Unit 789 is likely more accurate, because 100% seawater is ~56000 microsiemens (see figure). We recalibrated both using the Oakton handheld meter and saltwater dilutions.

Data from June 21st collection of conductivity measurements. 874 and 789 are the two Leveloggers used. They show differences in measurements taken, but same relative movement.


There were small temperature dips in the data where there was no dip in conductivity, and vice versa. As a whole, temperature and conductivity correlate in the water.

This graph shows that temperature and conductivity correlate. The large dips in conductivity is when the meter left the water.

The two maps do not show any trends that indicate point-source seeps. The conductivity spikes downward in front of stream outlets, as was expected. The general trend is that conductivity rises the further away from shore.  Confounding factors include the time of day the data was collected at. The maps were created in Arcmap 10.1.





Pore-water sample taken away from stream outlet.


Pore-water sample near stream outlet.


(c)  A Levelogger  was placed a few yards inland from where the perennial channel lets into the ocean. It shows that the conductivity level changes more with temperature than with tidal movement.  Over the two nights the Levelogger was there, there was only a minor change in conductivity. The sea water likely only goes that far inshore during storm events.

Conductivity is measured every 15 minutes.


Tidal Chart



There are a few explanations for the data. The most likely is that SGD is diffuse along the shoreline. Another possibility is that the SGD plume encompasses the entire study area and that is why there is so little variation in the data. SGD may also be further out in the ocean than the study area for this project. More research is needed to come to a conclusion about the nature of any SGD in the area.


For future study:

  • Use thermal imagery over a larger area. If the thermal imagery could be mosaiced, surface heat patterns would likely emerge from the water.
  • Chemical tracers have proved in other study sites to be a valuable tool for identifying SGD, including in the nearby Kahana Bay.2
  • Conductivity and temperature meters could be used over larger time scales and in consistent areas to give a better idea of shifting conditions in the ocean environment.



1. Burnett W.C. et al. “Quantifying submarine groundwater discharge in the coastal zone via multiple methods,” Science of the Total      

            Environment (2006): 1-46. Date accessed June 25, 2013, doi: 10.1016/j.scitotenv.2006.05.009

2. Garrison, G.H., Glenn C.R., and McMurtry G.M. (2003). Measurement  of  submarine groundwater  discharge  in  Kahana  Bay,  O’ahu,

Hawai’i. Limnol. Oceanogr., 48(2): 920-928

3. Johnson, A., et al. (2008) “Aerial infrared imaging reveals large nutrient-rich groundwater inputs to the ocean,” Geophysical

          Research Letters (35): 1-6. Date accessed June 25, 2013, doi: 10.1029/2008GL034574

4. Presto, M.K., et al., 2007. “Submarine Groundwater Discharge and seasonal trends along the coast of Kaloko-Honokohau National

Historic Park, Hawaii, Part I: Time-series measurements of currents, waves and water properties” U.S. Geological Survey Open-

        File Report. November 2005 – July 2006. U.S. Geological Survey Open-File Report 2007-1310.