Bear Encounters with Seismic Stations in Alaska and Northwestern Canada

Abstract

A typical seismic experiment involves installing 10–50 seismometers for 2–3 yr to record distant and local earthquakes, along with Earth’s ambient noise wavefield. The choice of the region is governed by scientific questions that may be addressed with newly recorded seismic data. In most experiments, not all stations record data for the full expected duration. Data lossmay arise from defective equipment, improperly installed equipment, vandalism or theft, inadequate power sources, environmental disruptions (e.g., snow covering solar panels and causing power outage), and many other reasons. In remote regions of Alaska and northwestern Canada, bears are a particular threat to seismic stations. Here, we document three recent projects (Southern Alaska Lithosphere and Mantle Observation Network, Fault Locations and Alaska Tectonics fromSeismicity, andMackenzieMountainsEarthScope Project) in which bears were regular visitors to remote seismic stations. For these projects, there were documented bear encounters at 56 out of 88 remote stations and 6 out of 85 nonremote stations. Considering bear-disrupted sites—such as dug-up cables or outages—there were 29 cases at remote stations and one case at nonremote stations.We also compile bear encounters with permanent stations within the Alaska Seismic Network, as well as stations of the Alaska Transportable Array. For these two networks, the stations are designed with fiberglass huts that house and protect equipment. Data losses at these networks because of bears are minor (<5%), though evidence suggests they are regularly visited by bears, and data disruptions are exclusively at remote stations. The primary goal of this study is to formally document the impacts of bears on seismic stations in Alaska and northwesternCanada.Wepropose that the threat of damage from bears to a station increases with the remoteness of the site and the density of bears, and it decreases with the strength and security of materials used. We suggest that low-power electric fences be considered for seismic stations—especially for temporary experiments —to protect the equipment and toprotect the bears.With the goal of 100%data returns, future seismic experiments in remote regions of bear country should carefully consider the impacts of bears. Supplemental Content: Evidence of bear encounters with seismic stations. INTRODUCTION Alaska is one of the world’s most seismically active regions. Its southern tectonic boundary is marked by the subduction of the Pacific plate under the North American plate. Toward southeastern Alaska, the plate boundary is marked by the Queen Charlotte and Fairweather strike-slip fault system. Over the past 100 yr, nearly the entire plate boundary has ruptured over the course of dozens of earthquakes, including the 1964 Mw 9.2 Prince William Sound subduction earthquake, the 1979 Mw 7.4 Wrangell St. Elias thrust earthquake, and the 1949 M s 8.1 Queen Charlotte strike-slip earthquake. Mainland Alaska and northwestern Canada—far from the plate boundary—exhibit broad crustal deformation evidenced from differential surface velocities, seismicity, and active faults (Page et al., 1991; Freymueller et al., 2008; Koehler and Carver, 2018). The subsurface structure comprises accreted terranes and, at deeper levels, is dominated by the subducting Pacific plate, which flattens under south-central Alaska in response to collision of the Yakutat block (Eberhart-Phillips et al., 2006; Christeson et al., 2010). Alaska has been a prime target for seismic experiments seeking to understand the structure and dynamics of subduction zones, starting with the Broadband Experiment Across the Alaska Range (BEAAR) in 1999–2001. This experiment, as well as those that followed (Table 1, Fig. 1), exploited the logistical advantage of Alaska’s limited road system. Recent experiments, such as Southern Alaska Lithosphere and Mantle Observation Network (SALMON) and Fault Locations and Alaska Tectonics from Seismicity (FLATS), ventured into remote regions of Alaska. Bears can be found nearly everywhere in mainland Alaska (Fig. 2). In this study, we document recent bear encounters with seismic stations in Alaska and northwestern Canada. Our approach is to document disruption at sites and to classify bear encounters. We augment this documentation with motiontriggered video recordings of animal activity at one site. Seismic experimental designs that had previously performed well along the Alaska road system were not adequate in remote regions, due to data losses caused by bears. Future seismic 1950 Seismological Research Letters Volume 90, Number 5 September/October 2019 doi: 10.1785/0220190081 Downloaded from https://pubs.geoscienceworld.org/ssa/srl/article-pdf/90/5/1950/4825210/srl-2019081.1.pdf by Kasey on 06 December 2019 experiments in remote regions of Alaska and Canada will need a strategy for deterring bears and reducing data losses. STUDY AREAS, SEISMIC NETWORKS, AND STATION DESIGN Starting in 1999, Alaska has hosted several regional seismic experiments funded by the National Science Foundation, as well as the state-wide EarthScope Transportable Array (TA) (Fig. 1). The three permanent monitoring networks in Alaska are Alaska Seismic Network (AK), AlaskaVolcano Observatory (AVO) seismic network (AV), and the National Tsunami Warning Center Alaska Seismic Network (AT). There are also isolated stations in mainland Alaska that are part of global (II.KDAK, IU.COLA) and national networks (US.EGAK, US.WRAK). SALMON (ZE) The SALMON experiment contained 28 stations in the Cook Inlet region of south-central Alaska. These stations required a range of approaches, including boat, float plane, all-terraine vehicle, and helicopter. The station design is described in Tape et al. (2017). Each station included a direct-burial Nanometrics T120 PH posthole sensor, powered by six Cegasa Celair 3V 1200 AH air alkaline batteries stored in a partially buried, 24 gal plastic box. No telemetry was used. The only adaptation following the first station servicing (in 2016) was to install all Global Positioning System (GPS) antennae inside the station box. In the first year, at least three GPS antenna cables (HLC1, HLC2, and BULG) were severed at sites where bears had caused station outages. FLATS (XV) The FLATS experiment contains 13 stations spanning the Minto Flats fault zone and Nenana basin, west of the town of Nenana in central Alaska (Tape, Silwal, et al., 2015; Tape et al., 2018). Most sites require boat access via the Tanana river. The FLATS experiment started in the fall of 2014 with the installation of two stations, FPAP, adjacent to a house near Nenana, and F3TN, a site 20 km downriver from Nenana. Figure 3a shows a typical site for the FLATS experiment. The seismometer is directly buried and connected to the station box via a sensor cable. The station box is a 53 gal (0:2 m3) Contico model 3725 Tuff-bin box. It contains twelve 3 V batteries, in two 18 V parallel sections, which power the sensor, a Kinemetrics Quanterra Q330 recording system, a GPS antenna, and a radio antenna. A key difference between FLATS and Table 1 Summary of Bear-Caused Data Outages at Seismic Networks in Alaska and Northwestern Canada Duration Stations Bear-Caused