Scientific Investigations Report 2006-5136
U.S. GEOLOGICAL SURVEY
Scientific Investigations Report 2006-5136
A self-contained temperature measuring and recording probe, encased in stainless steel and designed for ground-water monitoring, was selected as the best option for thermal profiling because no cables are needed to be towed, therefore, preventing damage to or loss of sensor/cable. The probe (CTD-Diver™, Van Essen Instruments; now manufactured by Solinst® and called a Levelogger®) has the added benefits of also measuring conductivity and depth with a variable sample rate. The probe (fig. 2) is 2.2-cm diameter, 26.0-cm long, and weighs 160 grams; an optical read-out unit connected to an LTC is used to program and read out the datalogger.
Probe accuracy was rated as 0.1°C for temperature, 0.1 percent of the full range of depth (in cm), and 0.05 µS/cm for conductivity. The reported accuracy of the probe’s internal clock is better than 1 second per day at 20°C. Temperature measurement reliability was checked using a controlled temperature bath (accuracy of 0.05°C) over a range 6 to 30°C. Results confirmed an accuracy of 0.1°C over the full range and temperature change over 1-second intervals was captured by the probe. A probe that only measures temperature or temperature and depth also can be used in this method.
A light, rugged container consisting of white 5.1 cm plastic pipe with a rounded cap on one end, and a total length of about 38 cm was designed for the probe (fig. 2). Water flowed through 1.5-cm holes drilled throughout the container body. A stainless steel eye bolt on the end of the cap with two bolts provides a connection for a 15-m tow rope. A 15.25-cm piece of 2.5-cm diameter flexible wire conduit is centered in the cap, and a bolt hole is drilled through (perpendicular to) the cap and conduit; a bolt allows for probe attachment. A greased bolt is then inserted and the surrounding area is filled with silicon caulk; greasing the bolt allows for easy bolt removal after the caulk hardens. Caulk helps prevent the bolts on the eyebolt from loosening, holds the conduit in place, and adds weight to the front of the container for towing. About 9 cm of the probe extends into the conduit and the probe extends about 2 cm inside the container end. A neoprene bicycle handle is inserted over the probe extending beyond the conduit (but not covering the sensor) and is held in place with 2.5-cm wide, 10-mil, all weather pipe-wrap tape. The conduit and neoprene prevent the probe from contacting the sides of the container, but allow some movement to absorb shocks when contacting various bottom structures such as boulders. Neoprene also provides some buoyancy needed to allow the container to move around or over a stream’s bottom structure such as large boulders. After many profiles, the container withstood constant contact with the streambed, boulders, woody debris, and other objects.
One or two probes can be used in this method. Two probes initially were used during method development; currently (2006) one probe is deployed for a profile. If using two probes, one is for streambed measurements and the other is for shallow measurements. If more than one reach is profiled using two probes, for consistency in inter-reach comparison, identify one probe for shallow measurements and the other probe for streambed measurements.
While developing and testing this method, the probe’s sample rate was set at 3 seconds. For conductivity measurements, Harvey and others (1997) used a sample rate of 0.1 sec and Lee and others (1997) used a sample rate of 0.05 sec. The 3-second rate was used based on expected streamflow velocity and initial reach length because measuring temperature every 0.2-6 m would capture differences representative of large discharge quantities without compromising probe storage capacity and allow for sensor equilibration. However, based on datalogger capacity and profiled reach length with its attendant average velocity, the sample rate (ranging from 1 to 3 seconds) was adjusted for each reach. Expected average velocity for a reach should be analyzed in conjunction with the reach length to estimate a sample rate based on storage capacity of the particular probe and the GPS used. Additionally, if possible, the probe and GPS should use the same sample rate to simplify merging the resulting two data series. A slower sampling rate can be used depending on the specific purpose for measuring a profile. For example, other investigators used this method along a reach with measured large quantities of ground-water discharge and set the sample rate at 6 seconds to identify the area where ground-water discharge begins (G. Gregory, written. commun., Washington State Department of Ecology, 2006).
Two methods were used to determine the location of the watercraft towing the CTD-Diver™. First, a Garmin 76 GPS was connected to a LTC using Garmin’s MapSource™ software. MapSource™ enables streaming of GPS data to a stored route at a 1-second sample rate. Each GPS data point on the route is time stamped, and latitude, longitude, altitude, and for each leg between readings, the length, speed, and course are recorded. For this method, the LTC was placed in a waterproof container and the GPS was placed in a smaller container bolted to the top of the LTC container and sealed with caulk. An opening between the two containers allowed for cable connection between the GPS and LTC. The LTC operated from a small gel-cell battery through an inverter that makes a warning sound when the battery is low. Two back-up batteries should be carried in waterproof bags.
Second, a Trimble® GeoXM™, operating with TerraSync™ and Geoexplorer CE™ software, was used. This method does not require an LTC in the watercraft, and the GeoXM™ collects GPS coordinates at a user-defined time interval and accuracy. Both GPS units are WAAS (Wide Area Augmentation System) enabled and when receiving the WAAS signal, the horizontal accuracy generally is less than 3 m (15 m without WAAS). The Garmin® set-up from the first method also was used to synchronize the clock in the probe to GPS time.
To start a profile, the LTC’s time is set to GPS time using the Garmin® and its software. The probe’s internal clock is then set to the LTC’s time, closely synchronizing it to GPS time. Start times and sample rate of the probe or probes are then set.
The probe or probes are then placed in the water for at least 5 minutes to equilibrate to the ambient water temperature. Equilibration is needed because the combined mass of the probe and container affects the measured temperature when a large temperature differential exists between the probe/container and streamflow. For example, on a warm day, the probe and the container may obtain an ambient temperature of 35°C during transport to a site and streamflow temperature may be 15°C. Starting a profile with such a differential would result in a sensor slowly equilibrating to streamflow temperature. After equilibrating to near ambient streamflow temperature, the semiconductor sensor is unaffected by the probe and container temperature.
Next, either the Garmin® GPS was connected to the LTC and a route was started with data logging or the GeoXM™ was started for data logging a line feature. A profile was then started with the shallow probe (if one is used) just under the water surface and the deep probe hand held with a towline. The deep probe is hand held for safety reasons. For example, if an attached tow line was snagged in fast current, the watercraft could become submerged or capsize. Although not done for this study, a small buoy can be attached to the towline. The buoy would allow towline recovery if it is lost. Profiling is as continuous as possible to avoid temporal discontinuities due to the water-temperature increases during the time the profiling is suspended. Profiles are best completed at lower flows, staying near the thalweg.
Last, temperature dataloggers were deployed at the head and tail of the profiled reach. Temperature data from these loggers provided additional information on the diurnal temperature change in water entering and leaving the reach. The difference between upstream temperature and thermal profile also provided information on variations from the expected warming of a water parcel. These differences relate to the change in the heating rate.
The type of watercraft used usually depends on physical characteristics of a particular reach and streamflow volume. Profiling was completed over a wide range of flows, about 1 to 110 m3/s. During the method’s initial development, a two-man inflatable raft with a rowing platform was used. To obtain greater mobility and control, a two-person, self-bailing inflatable kayak was used instead of the raft and was a good work platform. A motorized watercraft was used for one reach, but because it did not provide as much control as a kayak for measuring in a Lagrangian framework, it was determined that motorized watercraft is best suited for non-braided reaches with streamflow velocities less than about 0.2 m/s. Other investigators using this method have used a single-seat, inflatable pontoon watercraft (G. Gregory, Washington State Department of Ecology, written. commun., 2006), which also should be a good working platform.
Time required to profile a reach depends on reach length, average velocity, and potential impediments such as diversion dams or log jams. Profiling a reach of about a 5 km starts at about 10:00-11:00 a.m., and for longer reaches profiling generally starts as early as 8:30 a.m. to ensure that diurnal heating is strongly initiated. Ideally, reaches are profiled on cloudless days. Shorter reaches with lower streamflow volumes can be profiled by wading. A handheld GPS with data logging capabilities is the best instrument to use for waded reaches.
Profiling in a Lagrangian framework at velocities as much as 2.0 m/s results in the sensor moving rapidly over the streambed. In turn, low volumes of ground-water discharge results in a high volume of stream water contacting the sensor compared to ground water. Therefore, a high-resolution survey of temperature variations due to ground-water discharge at the streambed is not documented using this method. Testing the potential to conduct high-resolution surveys indicated the method would work, but the maximum drift velocity should be less than about 0.20-0.25 m/s. This conclusion however, is based on conditions during the profiling while developing and testing the thermal-profiling method. For neutral or losing parts of reaches, the thermal profile resolution generally is good and is a function of combined effects of the sampling interval and streamflow velocity.
Another limiting factor is that a particular volume of water measured may not be representative of streambed water in ground-water discharge areas due to mixing. Mixing can mask the ground-water signature and is affected by channel morphology and streamflow volume and velocity. Probe location relative to the streambed also varies. Problems with the actual location of measurements arise because the tow-line length can change (0.5-3 m) to keep the probe as close to the raft as possible for accurate GPS locations while attempting to stay at the streambed. Length also is changed when moving through rapids, areas of woody debris, and riprap or to avoid submerged objects that can snag the streambed probe. These tasks are difficult in fast-moving current and rapidly changing bottom conditions. Therefore, the accuracy of the measurement location is reduced, but the position is accurate when referenced to the nearby logged positions. Of the tens of thousands of measurements made during the method testing and use, relatively few were upstream of the previous recorded position and few were on the stream bank. More challenging is the loss of GPS reception, in which case, the latitude and longitude are linearly interpolated between GPS values that bracket that period and are then assigned to the CTD data. This method produces a straight line track, which may not represent that part of a reach. However, Geographic Information System (GIS) can be used to fit the track to the channel center to improve accuracy of the profile location.
The nature of ground-water discharge also can confound profiling results because discharge can occur as low-volume diffuse to high-volume spring, and older, and perhaps colder ground water near the thalweg compared to younger, perhaps warmer, ground water near channel edges (Modica and others, 1997).
A thermal profile documents a linear track of temperature and does not capture any three-dimensional aspects of the thermal regime in a reach. However, areas of interest are readily identified. These areas can be investigated in more detail using discharge measurements in conjunction with mini-piezometer measurements.
U.S.
Department of the Interior | U.S. Geological
Survey
Persistent URL: https://pubs.water.usgs.gov/sir20065136
Page Contact Information: Publications
Team
Page Last Modified:
Thursday, 01-Dec-2016 19:06:35 EST
