The photos are hard to view. Adult albatrosses with bloodied heads. Gray, night-time shots of mice preying on birds that can’t or won’t leave their nests. Beautiful, dead birds, killed by the “vampire” mice of Midway National Wildlife Refuge.
The problem seems to have begun in late 2015. According to the U.S. Fish and Wildlife Service, “refuge scientists and volunteers found open wounds on the backs, necks, and heads of nesting albatrosses in a small area on Sand Island, Midway Atoll. Using automatic cameras, live traps, and laboratory examination of bite marks on mortally wounded birds, scientists rapidly identified the cause of the injuries as attacking non-native house mice.”
Since then, the FWS says “more than 300 nesting adult albatrosses are known to have been attacked and the mouse aggression has spread to two additional areas… Fort
y-eight bitten birds are known to have died and at least 46 nests have been abandoned.”
To address this, the service has published an environmental assessment, with the proposed action being the aerial broadcast of a rodenticide, brodifacoum, in pellet form over all 1,128 acres of Sand Island, the largest of the small islets in the atoll. (The public comment period on the draft EA closed on April 20.) The expectation of the FWS is that within a year of implementation, non-native mice will be eradicated “for the benefit and protection of nesting albatross species (e.g., Laysan, short-tailed, and black- footed), other nesting seabirds (e.g., Bonin petrel), and their habitats.”
But will their benefit come at the expense of an even more threatened species, the Laysan duck? The agencies involved – including the Hawai‘i Department of Land and Natural Resources, the Fish and Wildlife Service, and the Papahanaumokuakea Marine National Monument – have enthusiastically supported the plan, including with their publicity the gruesome photos of the handsome albatross with mouse-inflicted injuries.
Yet Michelle Reynolds, who has studied Laysan ducks and worked extensively with them, has serious concerns about the mitigation measures proposed to protect that much smaller, and far more endangered, population. “[T]he mitigation plan as described in the DEA will not eliminate risks and poses additional and significant threats to a critically endangered species,” she says.
Loyal Mehrhoff, now retired but formerly field supervisor of the Fish and Wildlife Service‘s Pacific Islands Office and director of the U.S. Geological Survey’s Pacific Island Ecosystems Research Center, also has weighed in with critical comments. While he supports the goal of eradicating the mice, the “project will undoubtedly impact endangered Laysan ducks to at least some extent and those impacts could potentially rise to be very significant,” Mehrhoff says, submitting comments on his own behalf.
Few would disagree with the goal of eradicating mice. Rats were eradicated from the atoll more than 20 years ago using a combination of snap traps and brodifacoum in bait stations. Since then, the mouse population appears to have begun occupying ecological niches formerly occupied by the rats.
Another factor that might help explain the aggressive behavior of the mice has been the virtual eradication of an invasive weed, Verbesina encelioides, also known as golden crownbeard. For years, volunteers at Midway had spent their days pulling the weed. But, starting around 2012, a grant from the National Fish and Wildlife Foundation allowed the refuge to treat the weed with a mixture of glyphosate and another herbicide, Milestone.
According to the draft environmental assessment (DEA), from 2012 to 2017, the area of ground covered by this member of the aster family went from 50 percent to less than one percent.
The albatrosses moved in quickly to occupy the denuded land, as evidenced in a 2017 photo on the website of Island Conservation, the New Zealand-based organization that has helped the Fish and Wildlife Service with a number of projects intended to restore the natural balance in island ecosystems. These include the recent efforts to eradicate rats on Lehua island and Palmyra atoll.
The removal of Verbesina may have had an unintended consequence, however. Almost certainly, invertebrates and seeds from the Verbesina were important sources of food for the mice. With very little other vegetation planted in its place and a drought in 2015, the scene was set for the sudden appearance on the scene of the “vampire mice.”
Regardless of the factor or factors that may have prompted the devastating change in mouse behavior, one thing was clear by 2017: the mice had to go.
After removal of the rats, Midway was viewed increasingly as a suitable site for a second population of Laysan ducks. Before the arrival of humans, the ducks were widespread on all the Hawaiian islands. At the turn of the last century, in the early 1900s, only a dozen or so of them were counted on the 1,016-acre Laysan Island.
Despite years of protective efforts, by the turn of this century, the duck population on Laysan was still extremely vulnerable. Ecologists began to express concerns that all the Laysan ducks’ eggs were in one basket – i.e., Laysan island. Should a catastrophic storm, disease, or tsunami strike, the species could be wiped out. In 2004 and 2005, 42 of the birds were translocated to Midway’s Sand Island, to restore a second population, as a form of insurance for the species.
The reintroduction succeeded. By 2010, the population at Midway was thought to exceed that at Laysan. Best estimates now of the duck population at Midway put it at around 500 individuals, according to the draft EA, representing around half of the known global population. (Another small group of ducks was translocated to Kure; this population, however, has not grown.)
But, the draft EA acknowledges, the ducks are extremely vulnerable to the toxicant. “Initial tests at sites … where non-toxic bait piles were put out and monitored indicated that Laysan ducks would readily consume bait pellets,” the draft EA states. “Thus, there is a clear primary route of exposure to the rodenticide as it is assumed they would consume bait. Since the ducks also consume invertebrates, there is a likely secondary exposure. The consequence of that exposure is presumed to be substantial, and without mitigation, a large number of individual ducks present on the island during the eradication would very likely succumb to the toxic effects of the rodenticide.”
The draft EA goes on to state, “an individual duck would need to ingest only 5 bait pellets to receive a potential lethal dose of brodifacoum. For secondary exposure, an individual duck would need to ingest 1.5 oz. (42.6 g) of contaminated invertebrate prey, which would be 57.5 percent of a bird’s daily food intake…. Without mitigation measures, the worst-case scenario is that the entire population on Sand Island could be at risk of mortality.”
“[T]he only effective mitigation and minimization strategy is to prevent the exposure of the ducks to rodenticide either through live-capture and holding ducks on Sand Island or to capture and temporarily translocate the birds to another island such as Eastern Island [another small islet within Midway Atoll] until the risk period passes,” the DEA states. In light of the fact that the duck population on Midway “is globally significant for this species,” it goes on to say, “a robust minimization strategy would need to be in place prior to implementation” of the project, with the goal being “to ensure that Laysan ducks persist on [the atoll] after the mouse eradication.”
A ‘Step-Wise’ Release
That “robust” strategy outlined in the DEA involves capturing ducks and holding them for a month or so, by which time the bait pellets would have degraded. After that, the ducks would be released “in a step-wise progression,” that is, a few at a time, with the released animals being monitored to see if they suffer any ill effects as a result of either direct exposure to the rodenticide or indirect exposure, since the brodifacoum residues will likely persist “and will likely enter the Laysan duck food web (invertebrates), leading to multiple and repeated exposures over time.” In this way, “any uncertain or unexpected loss of ducks can be detected early and before a significant portion of the population would be put at risk.”
Just how long the “step-wise” release of ducks would last is not even hinted at in the environmental assessment. “The temporal exposure risk and consequence of exposure is difficult to quantify a priori but is likely to have high consequence to some individuals for a few to many months post-bait application. … Effectively, each release group is a sentinel for the next group of released animals, and through monitoring for survivorship and other indicators, the mitigation team can either continue with the release of ducks or halt the release and re-capture some individuals.”
To ensure that a “local population” of Laysan ducks survives the eradication project, starting in the spring of 2019, in advance of the planned broadcast of brodifacoum, 50 male and 50 female ducks are to be captured, transferred to Eastern Island, and held there in aviaries for eventual release. Other ducks are then to be captured, banded, have their flight feathers clipped, and removed to Eastern Island. “Capture efforts will continue on Sand Island throughout the bait application period and any ducks not captured for hold-release, subsequently exposed to the rodenticide, and demonstrating signs of toxicosis, would be captured and treated … by a veterinary professional to offset the negative effects of the rodenticide.”
Under the Endangered Species Act, the Fish and Wildlife Service has to determine that the proposal will have no “net negative impact” to the species. Over the long term, the removal of mice will likely benefit the ecosystem. “However, short-term adverse impacts are likely even with minimiza- tion and mitigation measures in place,” it states.
The DEA goes on to note that the founding population of Laysan ducks on Midway was small “and this species can reproduce quickly. … [I]n six years this population grew more than 15-fold to 661 birds…. [T]here is likely to be short-term adverse impacts to the population of ducks on Sand Island, but the population should recover quickly, and thus the action is not expected to have long-term adverse effects to the Pacific population.” In other words, the Fish and Wildlife Service assumes that however many ducks are killed in the process, the remaining ducks will breed at these same high levels. But in the years since the Midway population was established, conditions have changed in a way that does not favor the ducks. As Reynolds notes in her comments on the DEA, at the time of the release, 2004-2005, “Midway atoll was densely covered in Verbesina weeds with more numerous fresh water seeps. … The population has not increased at the rate it did after the translocation since 2008.”
Sheila Conant, one of the most respected authorities on Hawaiian birds, told Environment Hawai‘i that to her, “the ‘step-wise release’ protocol sounds like an experiment. ‘Hmmm, let’s release a few ducks and see what happens. If they don’t die from poisoning, let’s release some more!’ I would think that doing this kind of an experiment with the most robust popula- tion of a highly endangered species would be against the law.”
Reynolds wrote her dissertation on the foraging ecology, population dynamics, and habitat use of Laysan ducks. and assistance with the expansion of their range. She was also heavily involved in both the planning and execution of translocation efforts.
She supports the proposed action but is critical of the mitigation plan for the Laysan ducks. The “substantial risks to the critically endangered Laysan duck appear to be underestimated, and the proposed mitigation has important inadequacies, uncertainties, and feasibility concerns,” she writes in her comments on the DEA. “Broader and more effective mitigation actions are urgently needed to protect the Laysan duck, as the current plan substantially increases the species’ high risk of extinction and may reverse the recovery efforts of the last decade.”
The Fish and Wildlife Service has a recovery plan for the Laysan duck, which identifies the Midway population as crucial to the species recovery. The DEA does not consider any recovery actions in this plan, including “translocating a quantity of candidate at-risk birds off of Midway Atoll to Lisianski Island, Kure Atoll, and/or establishing a genetically managed captive breeding population” – all of which “would benefit the species,” Reynolds writes. Mehrhoff raises a similar point: “There are numerous options for minimizing impacts and mitigating/offsetting impacts. For example, the project could improve habitat for ducks, implement better anti-botulism efforts in the future, or bolster duck populations on other islands.”
These broader mitigation actions “would help preserve species genetics, reduce the probability of extinction due to random event and disasters, and is an opportunity to reduce the risks associated with toxicant and avian botulism exposure during and after the proposed toxicant application.”
While the “details, logistics, and justification of the broadcast operation are well described,” Reynolds says, “the period of secondary lethal and sub-lethal exposure is much less certain and could vary between 30 days and two years after the application. This secondary exposure risk to the ducks is apparently very high but not clearly mitigated. The plan to release endangered ducks and monitor them for toxicosis is incredibly risky. The mitigation actions, effectiveness monitoring, post-application actions, and contingency planning (should the effort fail) become less well considered and less complete.”
Reynolds also has concerns as to the timing of the application (in July 2019). “This is during the peak of [Laysan duck] breeding and molt, peak of seasonal avian botulism epizootics, and reduced food availability,” she notes. And the prospect of using the ducks as sentinels – releasing them a few at a time and observing whether they suffer as a result of exposure to the toxicant – is a huge risk, she adds. “The endangered birds should not be used as indicators of environ- mental toxin,” she writes, suggesting instead that carcass searches and non-endangered birds, such as mynahs, should be used as indicators of toxin in the food web.
The DEA glosses over the difficulties of holding Laysan ducks in captivity and is also inconsistent in describing the number of ducks proposed to be held during the bait drop, she notes.
Mehrhoff comments on the DEA’s lack of “key management triggers” that would protect the Laysan ducks and other species from unanticipated impacts. “For example, if Laysan ducks turn out to be very difficult to catch, what is the minimum number that must be captured for the project to proceed? … If ducks do poorly in captivity, what do you do? If hundreds of albatross chicks die after first application, what do you do? How many deaths does it take to stop or alter plans?”
An Alternative Toxicant?
The draft EA notes that a different type of rodenticide, AGRID (with the active ingredient of cholecalciferol) has been used in limited areas to control mice on Midway since 2016. The AGRID pellets are hand-broadcast under a supplemental label allowing its use for mouse control. Unlike brodifacoum, which is an anticoagulant, cholecalciferol interferes with the target animal’s calcium levels by increasing calcium absorption and reducing calcium excretion. The DEA states that AGRID’s effectiveness “has been proven in limited hand-broadcast situations, and it is relatively safe to non- target species if used according to label directions.”
“It should be noted that there were no observations of any non-target organisms such as shorebirds or Laysan ducks interact- ing with AGRID bait pellets in the field or being found sick or dead … as a result of the baiting process in 2016/2017,” according to the DEA. While there is a potential for the ducks to consume some bait, “to reach a lethal dose, a Laysan duck would need to ingest three times its body weight in pellets, which is unlikely to occur.”
Although AGRID is admittedly effective, the DEA does not seriously consider its use instead of brodifacoum. AGRID and other rodenticides that are not anticoagulants “are untested on islands larger than 22 hectares (54 acres). Furthermore, there is no cholecalciferol product registered by the EPA for aerial broadcast and the purpose of island-wide eradications for mice. Using [Midway] as a test island, without a high probability of success, would be inappropriate due to the high financial cost of the operation.”
The DEA sums up why the use of AGRID or other types of non-anticoagulant rodenticides were not further considered. There’s the high financial risk of using a bait that’s not “tested on islands comparable to Midway, potential bait avoidance, and greater human safety risk” – all of which “disqualifies them from detailed consideration for use” on Midway.
In addition, the DEA dismisses the notion that hand-broadcast of brodifacoum or the use of strategically placed bait stations could achieve the desired result.Bait stations would need to be deployed in a much denser concentration for mice than for rats, given the smaller home range of mice. “Island Conservation estimates that a minimum of 45,200 bait stations would be needed to cover the total area of Sand Island,” the DEA states. “[M]ore than approximately 280 miles of trails would need to be opened, flagged, and maintained to support crews walking to install, service, monitor, and decommission these stations. These trails would need to be opened in key habitat such as the coastal fringe in high density [naupaka] and through habitat with Bonin petrel burrows, which are found wherever the substrate allows for excavation by the birds… It is likely some burrows would be stepped on and collapse suffocating adults or young. Island Conservation estimates that, assuming a manageable crew size of 40 workers, this would require 200+ days, and an individual station would need to be visited at a minimum of five-day intervals.”
Mehrhoff takes exception to the discussion of alternatives, noting that the environmental impact statement prepared in advance of the rat eradication project at Palmyra “did a better job of explaining the rodenticide selection process.” The explanation of why diphacinone, a first-generation anticoagulant rodenticide, was not selected for use “is weak and needs to be expanded. … This is even more important at Midway, given the much higher potential for significant non-target mortality from brodifacoum compared to diphacinone to Laysan ducks and other species.”
— Patricia Tummons