It’s difficult to detect drywood termites because of their cryptic nature. A University of California researcher shares his data and offers PMPs best practices for using the Termatrac T3i.
The western drywood termite (Incisitermes minor) (see Figure 1) is an important structural pest, especially in California and Arizona, causing an estimated damage of $250 million annually (Cabrera and Scheffrahn 2017). These termites are hard to detect in structures due to their very cryptic nature. Most drywood termites live entirely within wood and rarely leave their galleries. The only time people notice their presence is when alates (winged termites) swarm, when fecal pellets are seen in or around the structure, or when inspections reveal subsurface tunnels within structural wood. Among these signs, only the presence of live alates is proof of an active infestation (the other two may only indicate old infestations).
Drywood termite control methods in structures are generally divided into local and whole-structure treatments. The efficacy of any local treatment is dependent on the accurate detection of active infestation sites and, sometimes, access to subsurface galleries. Therefore, researchers and pest management professionals have sought non-destructive methods and tools to pinpoint active infestations within structural wood members. Several different methods — such as x-ray, acoustic emission, low-energy microwaves, canine detection and infrared cameras — have been tested by researchers for their efficacy in detecting drywood termites. Some of these methods are currently utilized by PMPs, but based on published research, there is no “silver bullet” when it comes to drywood termite detection. However, when used properly, any of these methods can partly reveal infestations inside a structure, helping PMPs to better control drywood termites.
One relatively recent termite detection technology on the market uses radar (low-energy microwaves) to detect motion; it is commercially known as Termatrac. The first model, Termatrac T1r, came out in 1999. It was eventually followed in 2010 by the second and latest model, the Termatrac T3i (Taravati 2018). Termatrac T3i is different from acoustic emission devices, which detect sound created by termites biting and nibbling on wood (Scheffrahn et al. 1993). In addition to detecting motion within wood, Termatrac T3i can also detect moisture and measure surface temperatures, using two other sensors. One challenge when using termite detection devices is interpretation of their output. In this case, Termatrac’s radar sensor provides output in the form of a line graph as well as a bar graph (see Figure 2).
Like acoustic emission devices, Termatrac can pick up signals from non-target sources — such as moving vehicles, people, pets, other insects (e.g., ants), water in pipes, and strong shaking or vibrations. All of these may produce a false positive result, meaning the device shows an indication of live termites when no termites are actually present. Also, the output signal weakens as the depth of detection (distance from the sensor to a termite) increases; this may produce a false negative result (not detecting termites when they are present). Therefore, PMPs must be careful when interpreting Termtrac’s output graphs. In order to evaluate the termite detection efficacy of Termatrac T3i and to improve interpretation of its output, I conducted field and laboratory studies in Southern California. A technical account of the laboratory work has already been published (Taravati 2018,).
FIELD STUDY. Termatrac T3i was tested in five different single-family houses (see Table 1) in Southern California, most of which had confirmed drywood termite infestations in the past. Elements of structures inspected included interior and exterior walls, baseboards, wood fences, and frames of windows and doors. The device was tested using several operational methods:
During field inspections, residents were asked to minimize noise and movement (of both people and pets) in and around the house. Careful observations were made on the position of the Termatrac device, fluctuations in output signal, and the presence of any movement in front of, at the sides of, behind the device and on the other side of the wall.
RESULTS. Several sources of interference, leading to false positive results, were identified during this study. These included:
In general, using the Termatrac’s radar surface at 45 degrees to the inspection surface resulted in more interference as compared to when the radar’s surface was flush against the wall. Surprisingly, Termatrac did not pick up vibrations from heavy machinery working loudly nearby. This makes Termatrac an available inspection tool, even for operators in noisy locations.
LABORATORY STUDY. Since drywood termite detection attempts are difficult to evaluate in the field without using destructive methods (removal and cutting of wood members to confirm infestations), laboratory experiments were necessary to evaluate the efficacy of Termatrac. The objective of these experiments was to find a practical way to visually interpret Termatrac’s signal, helping PMPs to use the technology with more confidence and more accuracy. These experiments also aimed to measure the effects of wood type, termite density and detection depth on signal strength.
A drywall model (see Figure 6) was constructed using gypsum patching panel drywall (The Sheetrock Brand) painted with white acrylic paint and primer (Behr Premium Plus Ultra Stain-blocking Paint & Primer) and a standard 2 by 4 wooden wall stud. Two wood types were included in this study:
Each stud was cut into 1-, 3- and 5-cm pieces (see Figure 7) using a power saw. Detection experiments were then conducted in the laboratory using two different targets:
Signal strength was measured when using different combinations of wood type, wood thickness and presence of drywall. In experiment I, the source of movement was a single Hexbug Nano micro-robot buried halfway into sand in a plastic vial. The sand was used to limit the movement of Hexbug in the vial, decreasing signal intensity to be more similar to that produced by real drywood termites.
In experiment II, vials contained either one or ten live drywood termites. As in experiment I, different combinations of wall layers were used between the Termatrac device and the target. Both experiments were replicated 10 times. Hexbugs or live termites were placed on one side of the model drywall and the Termatrac device was placed on the opposite side. The device was turned on and connected via Bluetooth technology to a Samsung Galaxy Grand Prime smartphone running Termatrac software application (Ver. 3.19). The “radar” mode was selected and 15-20 seconds was given for the device to self-calibrate before reading the signal (this protocol was according to the manufacturer instructions). After self-calibration, a screenshot (see Figure 2) was taken of the device output; this image was later transferred to a PC for analysis. To be able to measure and compare Termatrac’s signal quantitatively, line graphs from the screen shots were processed in Adobe Photoshop CC (2017) to approximate the area under the curve, measured as the number of pixels squared.
RESULTS. Experiment I used two wood types and several different detection depths. Wood type did not have a significant effect on signal strength while detection depth certainly did. Signal strength decreased as the number and thickness of “wall” layers increased. The results from experiment II showed that termite density (either one or ten termites per vial) did not significantly affect signal strength. On the other hand, as with experiment I, “wall” layers significantly affected signal strength. Experiment II showed that Termatrac is able to detect a single drywood termite, confirming the manufacturer’s claim. I was able to detect a single drywood termite even when it was behind drywall and a 5-cm piece of wood (total “wall” thickness: 6.3 cm).
Termatrac screenshots can be used as general references for typical device signals in the presence of drywood termites in structures. Shapes and heights in the line graph output vary greatly between replicates, making it difficult to generalize the output signal. Also, weak signals coming from termites deep inside the wood makes it hard to distinguish between live termites and signal interference from other sources. Sometimes, this can be overcome by reducing the gain. However, lower sensitivity associated with lower gain values may reduce the detection range or overall ability within a structure.
PMP TIPS FOR USE. To conclude, laboratory results when using Termatrac T3i with simulated “wall” showed potential promise for drywood termite detection. Nevertheless, PMPs should bear in mind that real field situations will always be different from the laboratory settings used. The following suggestions are provided to help PMPs use Termatrac more efficiently and accurately:
The author is an urban IPM adviser at the University of California Cooperative Extension — Los Angeles County office.
Author’s Acknowledgments/Note: I would like to thank Dr. Andrew Sutherland; Dr. Darren Haver; Termatrac’s Rick Wakenigg, Michelle Lutz and Victoria Johnston; Dr. Mike Rust; Dr. Dong-Hwan Choe; Dr. Rachel Surls and Valerie Borel; all the UCCE-Los Angeles Master Gardeners who allowed me to use their property for testing. This project was made possible by funding provided by the statewide UC-Integrated Pest Management Program (UC-IPM) program.
Figure 6: Experimental setup in the laboratory, showing the Termatrac device, a painted section of drywall, wooden “studs,” and a vial containing sand and a Hexbug Nano taped to a coffee container partially filled with sand.
Figure 8: HEXBUG Nano micro-robots used in lab experiment I. These robots move in an erratic fashion.
Figure 9: Image processing using Photoshop CC on screenshots from Termatrac applications to calculate the area under the curve as a measure of signal strength. a) original radar line graph b) radar line graph after image processing.
Figure 10: Results from experiment I comparing signal strength among different treatments. Different lower-case letters within each wood type indicate significant differences among treatments. NoObs: No obstacle between Termatrac and HEXBUG; DW: Drywall; NoHexNoTerm: no HEXBUG or live termites were placed on the opposite side of drywall.
Figure 11: Results from experiment II comparing signal strength among different treatments. Different lower-case letters within each termite density level indicate significant differences among treatments. NoObs: No obstacle between Termatrac and HEXBUG; DW: Drywall; NoHexNoTerm: no HEXBUG or live termites were placed on the opposite side of drywall.
Figure 12: Termatrac’s signal output at maximum sensitivity (Gain: 10) from experiment II on two densities of live drywood termite and two ‘wall’ treatments showing five replicates. As can be seen, signal strength (height of lines) varies greately even among the outputs within the same column.