The Santa Ana Sucker (SAS) is listed as threatened under the U.S. Endangered Species Act. The Act protects and aims to recover endangered species and their habitats. To support this goal, USGS scientists are researching what is necessary to restore and sustain healthy fish populations.

Current USGS research on the SAS began in 2015 when the San Bernardino Valley Municipal Water District partnered with the USGS to conduct a native fish survey.   The goal of the survey was to estimate native fish species abundance within a chosen section of the Santa Ana River designated as critical habitat. In addition to the SAS, the Arroyo Chub (Gila orcutti) was also studied. This was done in support of the Upper Santa Ana River Habitat Conservation Plan.

At the start of the survey, sampling methods included beach seining (netting) and snorkeling. Now backpack electrofishing is used exclusively. However, seasonal sampling constraints related to the fish's spawning cycle, as well as inconsistencies in methods, resulted in a limited understanding of fish population dynamics. As a result, there was a need for alternative, less intrusive sampling methods.

Recently, USGS researchers investigated the potential for passive integrated transponder tagging (or “PIT tagging”) to be used to identify individual fish and track their movements.  

PIT tagging was developed in the 1980s by NOAA Fisheries to investigate large scale movement patterns of salmon and steelhead. A pit tag consists of a microchip, a chip capacitor, and an antenna. These are all encased in a small glass capsule. Each tag has a unique code with letters and numbers that are used to identify individual animals. Tags are powered by a magnetic field and are recoded when they come within range of an antennae.
 

Media

Sources/Usage: Some content may have restrictions. View Media Details

The above illustration shows an antenna array generating an electromagnetic field (white lines), while a PIT tag inside the trout re-emits the array’s signal (black lines) as it passes through the field (Keefe et al., Forests 2019, 10(5), 458).

Study Design

In the late summer and early fall of 2024, scientists from the USGS visited the Riverside Corona Resource Conservation District (RCD), located in Southern California, and studied the effectiveness and safety of the PIT tagging sampling method. This work was funded by a California Department of Fish and Wildlife grant received from the U.S. Fish and Wildlife Service Cooperative Endangered Species Conservation Fund.

Media

Sources/Usage: Public Domain. View Media Details

The above photo shows members of the research team. From left to right: Tevin Bui (RCD), Jordan Buxton (USGS), Brett Mills (RCD), Alicia Ruan (RCD), and Brock Huntsman (USGS). Inset at right are Marissa Wulff (top) and Jeff Gronemyer (bottom), both of the USGS. 

Researchers, working at a facility of the RCD, used three recirculating artificial streams, or “raceways,” about 100 feet long, to simulate the SAS’s natural environment. Each raceway was divided into ten sections.

Media

Sources/Usage: Public Domain. View Media Details

The above photos show raceways at the RCD facility.

Scientists then selected 61 fish of at least 60 millimeters in length, into three groups:

1. Control group with no tag
2. Group with 8 mm tags
3. Group with 12 mm tags

Media

Sources/Usage: Public Domain. View Media Details

The above photo shows 8- and 12-millimeter sized PIT tags.

Fish in all groups were anesthetized and fish in groups B and C were tagged with either an 8- or 12-millimeter PIT tag. All fish were weighed and measured, then given 12 hours to recover before being released into the raceways.

Fish from each group were evenly distributed across the three raceway sections and monitored for 36 days for survival, tag loss, and growth. The survival rate during the study was 88%, with all deaths occurring within 48 hours of tagging. This survival rate is high compared to other studies where species of a similar size were tagged. At the end of the study, the raceways were drained, and the fish were collected, scanned, weighed, and measured. The tagged fish showed no difference in growth compared to the control group and only two fish lost their tags.

Results of the three methods used

The chart below shows the effectiveness of the three methods used. The black dashed line shows the true (or total) number of fish. The yellow line shows the number of fish the PIT tag reader was able to detect. The red line shows passive camera detections, and the blue line shows active camera detections. The closer the colored lines are to the true number of fish (black line) equates to the accuracy of the method. The PIT tag reader was the most effective method for detecting the number of fish in the greenbelt.

Media

Sources/Usage: Public Domain. View Media Details

Combining methods in the Santa Ana River for maximum effectiveness

Scientists then combined the two most effective methods: the PIT tag reader and the active camera. The combination was tested in the Santa Ana River. A PIT tag reader, along with two cameras were mounted to the bottom of a raft. At the same time, lures (or artificial fish) were outfitted with PIT tags and placed in the river. The raft then cruised a section of the river and recorded video while detecting the tagged lures. 

Media

Sources/Usage: Public Domain. View Media Details

The photo above shows (left) the raft outfitted with the PIT tag reader and two cameras, both facing down into the river. And (right) a PIT tag attached to a lure.

Using algorithm to identify fish

An open-source object detection system known as YOLO (You Only Look Once) was used to detect the fish and lures. The program algorithm, a simple form of artificial intelligence or “AI,” was “trained” to recognize the fish and lures by having been exposed to over 1000 images of fish and images that did not include fish.

The camera from the raft captured the above video.  As the camera comes in proximity to objects, lures are identified with LURE (light blue), and real fish are identified with SAS (dark blue).

In this pilot study, all the lures were detected, and the YOLO model was able to accurately identify and detect wild SAS. It’s hoped that this method can be used in an actual field study in 2025.

Further study of PIT tagging, and other methods used in this project is needed. With refinement of the AI-like model, it’s hoped that this method can be used more widely, in a greater variety of settings. Such tools will help identify individual fish, track their movements, and expand sampling, both spatially and temporally, enhancing the understanding of fish population dynamics.