Drawing attention to invasive Lymantria dispar dispar spongy moth outbreaks as an important, neglected issue in wild animal welfare

Summary

One aim of wild animal welfare research is to identify situations where large numbers of wild animals are managed by humans in ways that have significant welfare impacts. If the number of individuals is large and the welfare impacts significant, the issue is important. As humans are managing these animals, it is possible the welfare impacts could be moderated to reduce their suffering. The massive scale of invasive (e.g., non-native) Lymantria dispar dispar (spongy moth) outbreaks represents an unappreciated wild animal welfare issue, and thus deserves further attention from a welfare (not simply an invasive species-control) perspective.

The spongy moth is not endemic to North America. The species experiences localized three year-long outbreaks of half a billion or more caterpillars/km² every 10-15 years in regions where they are well established (including their native range). Spongy moths currently occupy at least 860,000 km² in North America, only ¼ of their possible range (though most of the occupied area is not experiencing outbreak conditions, most of the time). L. dispar continues to spread slowly to new areas each year despite multi-million dollar efforts to stop expansion. Assuming spongy moth caterpillars are sentient, methods for actively controlling them during outbreaks cause substantial suffering. The aerial spray (Btk) ruptures the stomach, causing the insect to die from either starvation or sepsis over two to seven days. However, because outbreaks are so large, most caterpillars are not actively targeted for control, and ‘natural forces’ are allowed to reduce the outbreak. The most prominent natural forces to rein in an outbreak are starvation and disease. The accidentally introduced fungus, Entomophaga (meaning “insect eater”) maimaiga, digests caterpillars’ insides before pushing through the exoskeleton to release spores, usually within a week. LdNPV virus is also common in the spongy moth population, but only causes high levels of mortality during outbreaks when larvae are stressed from extreme competition. A symptom of severe LdNPV infection is “larval melting,” or the liquefaction of the insect’s internal organs.

The scale of spongy moth outbreaks is tremendous, though notably these outbreaks are not necessarily higher-density than numbers of other insect species (e.g., 740 million to 6.2 billion individual wireworms/km2; Smithsonian, n.d.). However, spongy moths are one of the best tracked non-native insects (Grayson & Johnson, 2018; e.g., Stop the Spread program), providing us with better data for analyzing the scale of the welfare issue (both in terms of caterpillar density within outbreaks, and the total area affected by outbreaks). In addition, there is potential for significant range expansion by spongy moths that would increase the scope of the welfare concern, and there appears to be extreme suffering¹ induced by both active and natural outbreak control. As a result, spongy moth welfare during outbreaks could be an issue of concern for animal welfare advocates. Further research could improve spongy moth welfare by: 1) identifying the most promising long-term interventions for preventing/reducing the occurrence of outbreaks behind the invasion front, 2) contributing to halting the spread of spongy moths into new areas, and 3) identifying the highest-welfare outbreak management strategies where outbreaks do occur.

Introduction

A note on the welfare concerns of insects

The purpose of this report is not to determine whether insects are deserving of moral consideration as welfare subjects. Rethink Priorities has covered insect sentience and capacity for welfare in multiple posts that can be found here. A foundational assumption of this report is that the probability of insect sentience is non-negligible (see Gibbons et al. 2022), and therefore it’s appropriate to approach insects as welfare subjects.

Figure 1: Progressive spread of gypsy moth (L. dispar) across NE US from 1900-2007; compiled from county data by US Forest Service. Public Domain. Found at: https://commons.wikimedia.org/wiki/File:Gypsy_moth_spread_1900-2007.gif

Ecology of Lymantria dispar dispar

Lymantria dispar dispar (Lepidoptera: Lymantriidae)— referred to as the spongy moth in this report—(Entomological Society of America, 2021; Invasive Species Centre, 2022) is native to Europe. Spongy moths are currently established in the Northeastern and Midwestern regions of the United States, and in Eastern Canada (Invasive Species Centre, 2022; US Forest Service, 2022a) after an accidental release in 1868-1869 in Massachusetts (Forbush & Fernald, 1896). Spongy moths are often present at low levels in the environment, but their status as a significant pest arises due to massive outbreaks of millions of caterpillars/km² that occur every 10-15 years wherever the species is well established (for example, in New York State; NYS Dept. of Environmental Conservation, n.d.). Outbreaks may also occur, for similar reasons, in the moth’s native range (comparable data on these outbreaks were harder to find, and so we have focused this report on invasive spongy moth outbreaks).

The spongy moth has spread to over 860,000 km² in North America (Figure 1) and inhabits just one quarter of its possible range (US Forest Service, 2022b). Spongy moths are considered an ecologically and economically important invasive (non-native) species; the larvae voraciously devour the leaves/needles of over 300 species of trees and shrubs (Tobin et al., 2012). A single larvae can consume one square meter of foliage across development (Lallemand Inc./Bioforest, 2021). An outbreak can easily defoliate 100% of the host plants in a localized area for several years in a row, either killing them or seriously damaging them. In addition, large outbreaks can cause unpleasant rashes, upper respiratory irritation, and entomophobic reactions for nearby humans (Robinson, 2021; Tobin et al., 2012), and there may plausibly be unstudied physical irritation for nonhuman animals.

The spongy moth invasion front moved at a pace of ~20 km/yr until 1988 due to their generally uncontrolled, stratified dispersal, but the rate has slowed to ~4-10 km/yr since the start of the USDA “Stop the Spread” program, saving ~400,000 km² from infestation between 2000-2010 alone (Tobin et al., 2012; US Forest Service, 2022b). While spongy moth larvae (caterpillars) can only crawl ~100 ft before pupation (US Forest Service, 2022b), ballooning (windborne dispersal on silken threads) can transport larvae, during extreme weather, up to 190 km (Frank et al., 2013). Unintentional anthropogenic movement of egg masses often facilitates spread to new regions (such as highly localized populations found in the Pacific Northwest; US Forest Service, 2022b).

Spongy moths lay multi-layered egg masses containing a minimum of 100, but usually 300-500, eggs (Tobin et al., 2012), though larger egg masses of 500-1,000 eggs are frequently reported during outbreaks (Collins, n.d.; Lallemand Inc./Bioforest, 2021; McCullough, 2022a). Spongy moths have only one generation per year, so eggs laid in the summer remain unhatched over the winter. Eggs hatch in the spring and larvae progress through five to six larval stages, depending on sex. Adults consume liquid diets and have short lifespans. Females of the European subspecies established in North America do not fly, and thus, outbreaks often remain fairly localized as females lay eggs nearby where they emerged. Cyclical spongy moth outbreaks are thought to occur in response to limited rainfall and average to above-average temperatures during the larval development period in spring. These conditions reduce fungal pathogen loads substantially (see ‘non-anthropogenic control’) and cause a multi-year population boom in areas with susceptible forest.

Outbreaks generally follow a three-year cycle (though they can be slightly longer). In year one, resources are abundant and control methods are low, causing a huge increase in the number of larvae that make it to adulthood, dramatically increasing the number of eggs laid. As a result of abundant resources and low rates of disease, the majority of spongy moths in year one of an outbreak may not have overall negative lives. In year two (the “peak” year), there is an explosion in the number of larvae emerging from the large number of overwintering eggs. These larvae will face extremely stiff competition for waning resources, increased anthropogenic control, and high rates of mortality due to density-dependent regulators like viral diseases. However, because of the large number of larvae, many will still make it to adulthood to lay eggs. In year three (the “decline” year), larvae will again face competition, as well as anthropogenic and natural control. Both years two and three can be expected to produce suffering for the insects involved in the outbreak.

Methods of active, anthropogenic control

The first North American outbreak of spongy moths occurred in the 1890s. The original, unsuccessful eradication effort cost ~37 million in (current, 2023) US dollars, despite covering only 2,500 km². Millions of dollars are spent just by federal and state governments in the US each year on management, mostly to stop the spread of the moth into new areas (Tobin et al., 2012). Private landowners and state/province, federal, and local governments may all participate in suppression efforts during outbreaks. Methods that have been employed in controlling spongy moths are: chemical/biological pesticides, horticultural oils, mating disruptors, growth regulators, manual methods (egg mass removal, barrier bands on trees, etc.), and parasitoid/pathogen introduction (NYS Dept. of Environmental Conservation, n.d.). Different methods may be deployed depending on the stage of infestation. For example, mating disruptors² are only effective where population density is low, and are frequently used at the invasion front (US Forest Service, 2022b) – they have been applied to 1.6 million km² by the US Forest Service alone since 2010 (US Forest Service, 2022c).

The most common method currently employed to actively suppress outbreaks is a Lepidopteran-specific (e.g., targets only butterfly and moth species) biological pesticide called Btk (Bacillus thuringiensis kurstaki) applied as an aerial spray (University of Wisconsin-Madison, n.d.). While more research is needed to completely understand the insecticidal nature of Btk, current research suggests that Btk toxins are released in the insect’s stomach when the caterpillar consumes sprayed plant material. The toxins rupture cells lining the stomach, causing an infection of the body cavity. The insect is unable to eat, and subsequently dies from either starvation or the infection within two to seven days. If used twice and timed appropriately, Btk kills 99% of caterpillars in treated areas (though some report reduced efficacy in the field; New Jersey Dept. of Agriculture, n.d.). Btk is typically used to target younger caterpillars (McCullough, 2022a; Toronto and Region Conservation Authority, n.d.; University of Wisconsin-Madison, n.d.).

Gypcheck is another aerially sprayed control method. Gypcheck is species-specific (US Forest Service, 2022d), and thus, may be used in cases where endangered Lepidopteran larvae may be present; it is less effective than Btk (only 24-67%) and works by infecting larvae with the LdNPV virus (see “non-anthropogenic control”; Webb et al., 2004).

Non-anthropogenic control

Natural factors are considered essential in preventing outbreaks and reducing them where they occur (as outbreaks frequently cover too many km² to be reasonably treated). Starvation (from severely defoliated and dying trees, paired with high conspecific competition) and fungal/viral pathogens are the most common factors involved in spongy moth regulation. Together, these factors are responsible for reducing 80-90% of an outbreak (at least, as reported by the New Jersey Department of Agriculture; New Jersey Dept. of Agriculture, n.d.).

The main fungus naturally controlling spongy moths is Entomophaga (meaning “insect eater”) maimaiga. E. maimaiga was introduced to North America in the 1980-90s and regulates spongy moths in a density-independent manner. E. maimaiga’s success as a control agent is highly dependent on the weather; wet and warm conditions during spring are important for E. maimaiga’s success in controlling spongy moth populations (McCullough, 2022b). E. maimaiga eats away at the insides of a caterpillar before exploding out through the exoskeleton to release spores. Four to five cycles of infectivity can occur each spring, so E. maimaiga likely takes a few days to a week to kill each individual. Spores remain in the soil overwinter in order to reinfect the next spring’s population. E. maimaiga is relatively species-specific, but can affect other Lepidoptera.

The main virus infecting spongy moths is a species-specific nucleopolyhedrovirus, LdNPV. LdNPV is a density-dependent regulator of spongy moth larval populations, and generally only causes high mortality when larvae are at high densities. High densities increase larval competition and stress, which is thought to make them more susceptible to the virus (McCullough, 2022b). The main symptom of LdNPV is ‘larval melting,’ or the liquefaction of the larval organs, which begins prior to death. High doses of LdNPV caused death within five days during lab trials, but lab conditions may not accurately reflect field doses and thus, time-to-death (Chen et al., 1998).

Egg parasitoids may also play some role in reducing the number of emerging spongy moths, however they can only reach the outermost layer of an egg mass with their ovipositor (meaning a maximum reduction of ~25% emergence). Endoparasites, such as the Tachinid fly Compsilura concinnata, were introduced to control spongy moth larvae in the 20th century, but only kill 5% or less of the larvae in an outbreak (Mahr, 1999). These flies consume host larval tissue from the inside before pupating and can have three to four generations per year. They must overwinter inside the larvae of other host species; as generalists, they have been found in 150 native species of American moths and butterflies (Arnaud, 1978; Strazanac et al., 2001).

Finally, severe cold during winter can reduce the subsequent spring’s spongy moth larvae numbers (generally, by killing egg masses above the snow line; Andresen et al., 2001). Multiple consecutive days with temperatures lower than -20^oF (-28.9^oC) will generally kill overwintering eggs (University of Illinois Extension, n.d.). Egg hatch rates reached 81.9% in more ideal conditions (with low solar load; Andresen et al., 2001), and somewhere between a 50% and 85% egg hatch rate might be a reasonable estimate in the absence of better data. As global temperatures continue to rise and freezing conditions become rarer, this method of control is likely to become less common; this may also facilitate the species’ range expansion.

Non-target animals

Competitors will likely face exceedingly stiff competition for resources during spongy moth outbreaks. Many of these competitors are likely to starve to death, just as many of the L. dispar starve to death due to intraspecific competition, during outbreaks. In addition, L. dispar outbreaks may lead to interspecific pathogen spillover and can also increase the number of generalist parasitoids (such as the C. concinnata), increasing parasite-related suffering in closely-related species.

While predators may immediately benefit from L. dispar outbreaks, due to the massively increased resource availability, the ecological consequences of outbreaks and their subsequent collapse are unlikely to be positive overall for individuals of other species. Predators that have more offspring due to the increased number of resources in the peak years may find those offspring starving in subsequent years, when both spongy moth and competitor species’ abundance have been massively reduced.

Beyond the loss of spongy moths and their competitors, reductions in tree health/survival following the outbreak will also decimate the resources available to wildlife. As just one example, healthy oak trees, which are one of the preferred food sources of spongy moth larvae, produce acorns that feed over 100 vertebrate animal species (Ober, 2020). Following a heavy spongy moth outbreak in the 1980s, acorn production dropped from 30,000-41,000 kg/km² down to < 2,200 kg/km² (270-370 lbs/acre down to < 20 lbs/acre) in affected regions of Maryland and Virginia; even moderate defoliation produced a ~50% reduction in acorns (Gottschalk, 1990). While most hardwood trees survive a few years of extreme defoliation in a weakened state, many conifers (which also provide shelter and food to many birds and rodents, and even emergency food to deer; Yamasaki, 2003) will die following severe defoliation in only one year. Depending on forest characteristics, overall tree mortality can easily approach 20% (40-70% for conifers) with rapid increases in the second year of defoliation (Davidson et al., 1999, 2001). Lastly, loss of canopy cover due to defoliation can cause other knock-on ecosystem effects by increasing the prevalence of inedible non-native plant species, removing habitat for nesting birds, changing ground temperatures/moisture levels (altering the plant and invertebrate species able to inhabit that area), and more. These are just a few of the possible, immediate negative effects facing individuals from other species that live through a spongy moth outbreak.

When spongy moth outbreaks are actively controlled, individuals from non-target species may be harmed by chemical or biological pesticides: for example, active control with Btk is not used in regions with endangered Lepidopterans due to non-target effects (Gypcheck is used instead). Chemical pesticides are infrequently used, although DEET was applied in numerous states in the 20th century in an attempt to control the spread, and can deter larval feeding (Sanford et al., 2014).

Roughly estimating North American L. dispar outbreak scale

When estimating the scale of L. dispar outbreaks as an animal welfare problem, we followed these assumptions:

1. We are not interested in estimating the number of all L. dispar caterpillars alive in North America across all invaded acreage. Of course, there are many caterpillars alive in non-outbreak, low-density conditions at all times; these animals are, presumably, experiencing similar welfare concerns as other wild caterpillars and do not represent the special case of outbreak conditions we are considering in this post.
2. We are not interested in estimating the number of L. dispar caterpillars in outbreaks that are not experiencing an unusual degree of suffering. Conditions in year one of the outbreak cycle are not likely to be welfare-limiting (e.g., resources are likely abundant, and disease/parasitism low, leading to the population increase which causes the outbreak). Suffering is expected to be much greater in years two and three of an outbreak, as described previously. We therefore estimate only the number of caterpillars in years two and three of an outbreak, to limit our numbers to those caterpillars experiencing an unusual degree of suffering due to outbreak conditions.

To roughly estimate spongy moth population numbers during these years, we considered: the number of egg masses/km² (typical method used to survey spongy moth population size), probable egg mass size, probable egg hatch rates, and the number of km² affected at a particular severity of infestation. We begin by considering a case study from the town of Pelham, Canada, where a large amount of useful data on these variables were collected and published.

Case Study in the Town of Pelham, Canada in 2020 and 2021

Pelham is a town of ~17,000 people in Ontario, Canada that occupies 126.43 km², or 31,000 acres. After complaints of severe defoliation by residents in 2017 and 2018, leading up to a 2019 outbreak, they treated some spongy-moth-affected acreage. They surveyed the town and surrounding rural area to determine the prevalence of egg masses/km² in both 2019 and 2020 (the eggs that would become caterpillars in outbreak years two [2020] and three [2021]). A total of 133 plots were surveyed, and 30% of egg masses were considered to be large in 2020, compared to 84% of 2019 egg masses (the end of the first year of the outbreak).

The guesstimate model for our estimation is available here. We used the data on the number and size of egg masses at surveyed sites in 2019 and 2020 to estimate the number of hatching caterpillars in Pelham in 2020 and 2021 (the crest and decline years of the outbreak, where suffering is more likely). Like the Pelham report, we assume a large egg mass has 501-1,000 eggs, while a small egg mass has 300-500 eggs. For ease of calculations, we assume the same distribution of large and small egg masses at all sites, regardless of the severity of the infestation at that site. In 2019 this distribution was 84% large and 16% small, in 2020 it was 30% large and 70% small. We assume a low potential hatching rate of 50% and a high potential hatching rate of 85% (see “non-anthropogenic control”).

We estimated the total number of caterpillars hatching from a number of egg masses per km² for each year (Table 1) using the following calculations:

Total # of eggs/km² = (# egg masses size of small egg mass % egg masses that size) + (# egg masses size of large egg mass % egg masses that size)

Lower caterpillar/km² estimate = Total # of eggs/km² 0.5

Higher caterpillar/km² estimate = Total # of eggs/km² 0.85

Table 1: Total egg masses/km² and caterpillars/km² produced by those eggs in outbreaks year 2 and 3, in each defoliation category^✳ taken from the guesstimate model^†

Defoliation category	2019 eggs (2020 caterpillars - year 2)			2020 eggs (2021 caterpillars - year 3)
	% sites in category	Mean egg masses/km2	Caterpillars/km2	% sites in category	Mean egg masses/km2	Caterpillars/km2
Nil	8%	0	0	4%	0	0
Light	18%	35 million	18 million to 30 million	21%	21 million	11 million to 18 million
Moderate	13%	190 million	94 million to 160 million	11%	110 million	53 million to 90 million
Heavy	4%	290 million	140 million to 240 million	7%	190 million	96 million to 160 million
Severe	57%	2.8 billion	1.4 billion to 2.4 billion	57%	1.3 billion	660 million to 1.1 billion

^✳ Defoliation categories were provided in the Pelham report: Nil – 0 egg masses/km²; Light – 1-125,000 egg masses/km²; Moderate – 125,001-375,000 egg masses/km²; Heavy – 375,001-500,000 egg masses/km²; Severe – >500,001 egg masses/km²

^† Note that the guesstimate model resamples each time it is opened so you may see slightly different figures than those presented here

To get the total number of caterpillars in each defoliation category, we multiply these figures by the area affected. We take the assumption that the defoliation profile of the surveyed sites reflects that of the entire town of Pelham³:

km² affected in category = (126.43km² [area of Pelham] % sites in defoliation category)

Lower total caterpillars in category = Lower caterpillar/km² estimate km² affected in category

Higher total caterpillars in category = Higher caterpillar/km² estimate * km² affected in category

This means that the total number of hatching caterpillars in the Town of Pelham was 100 - 180 billion in 2020 (peak, year 2) and 50 - 85 billion in 2021 (decline, year 3; Table 2).

Table 2: Estimated total number of caterpillars in outbreak years 2 and 3 in Pelham based on the guesstimate model

Defoliation category	2019 eggs (2020 caterpillars)		2020 eggs (2021 caterpillars)
	km2 affected	Total caterpillars	km2 affected	Total caterpillars
Nil	10.5	0	4.8	0
Light	22.8	400 million to 690 million	26.6	280 million to 480 million
Moderate	16.2	1.5 billion to 2.6 billion	14.3	750 million to 1.3 billion
Heavy	4.75	680 million to 1.2 billion	8.6	820 million to 1.4 billion
Severe	72.2	100 billion to 170 billion	72.2	48 billion to 82 billion
Total (from the model)		100 billion to 180 billion		50 billion to 85 billion

Therefore, the total number of caterpillars in Pelham in years 2 and 3 of the outbreak is approximately 150 billion to 260 billion caterpillars (with a lowest estimate of 44 billion and a highest estimate of 700 billion⁴).

Approximately 1% of caterpillars were treated (1.2 km² out of 126.43 km²) by public entities (private treatment area unknown); another 80-95% will die via pathogens, parasitoids, and starvation to reduce the population to pre-outbreak levels.

The accuracy of this estimate is affected by the validity of the previously mentioned assumptions about egg hatch rate, the number/distribution of eggs/egg mass, and the number of km² in Pelham affected at varying levels of severity. In addition, the town notes in its report that many of the surveyed plots were on the edges of forested regions; these plots may not accurately reflect the number of egg masses of more interior plots. Finally, the town sprayed 1.2 km² of land in 2020, and note that all plots on this land saw a decrease in the number/size of egg masses; depending on the number of plots, this treatment may disproportionately affect our assumptions about egg masses. This variability is reflected in the final numbers - but even in the most conservative case, the scale of the problem in just two years, in an outbreak within just one Canadian town, is tremendous.

Expanding out from Pelham

The province of Ontario, Canada saw moderate-to-severe defoliation of 5,864 km² by spongy moths in 2020 (Lallemand Inc./Bioforest, 2021), and 17,000 km² in 2021 (Pembroke Today, 2021), with no breakdown of how much area is moderate/heavy/severe. With an outbreak profile identical to the Pelham outbreak in 2020 and 2021,⁵ 6.5-11 trillion caterpillars in 2020 and 9-15 trillion caterpillars in 2021 were involved.⁶ See the guesstimate model for Ontario here.

The 2021 outbreak is expected to have defoliated up to 40,000 km² across the Northeastern United States and is one of the worst outbreaks in recent history. Previous records include a 1981 outbreak which defoliated 53,000 km² in one year (an area slightly larger than Vermont and New Hampshire combined) and a 1990s outbreak that peaked at 29,000 defoliated km² in a year (Aouga & Simko-Bednarski, 2021; McManus et al., n.d.; Figure 2). In between these larger outbreaks, small, localized outbreaks can still occur (see Figures 20-30 McManus et al., n.d.); for example, another recent outbreak in 2016-2018 affected 5,000 km² in eastern Connecticut (The Connecticut Agricultural Experiment Station, 2021).

A total of 350,000 km² or over 86 million acres of land have been defoliated in the United States due to L. dispar outbreaks between 1970 and 2019 (US Forest Service, 2022e). The severity of these outbreaks is not listed, but it is likely the defoliation was heavy to severe, making 0.1-1.4 billion caterpillars/km² a good range. This means 35 trillion to 385 trillion spongy moth caterpillars suffered through outbreak conditions in the Northeastern US in the last 50 years (0.7-7.7 trillion caterpillars/year).

For context, the United States and Canada together rear an estimated 68.6 to 102 billion insects for food and feed each year (Rowe, 2020). If 10 quintillion insects are alive on the planet at any given time (Smithsonian, n.d.), during the early days of US spongy moth outbreak season prior to mass mortality, one in every 1.3 to 14.3 million of the insects alive on the entire planet is a US-based spongy moth larva in an outbreak. At 10 mm in size (early-mid larval stages), a chain of the U.S. based spongy moth larvae experiencing outbreak conditions would stretch to the moon and back between nine and 100 times each year.

Future geographic range and outbreak potential

Spongy moths occupy only a quarter of their possible geographic range in North America. Because weather patterns differ in these parts of the spongy moth range, infestation frequency and severity may vary. However, it’s safe to assume that the problem will grow significantly if the range of spongy moths continues to expand. Northern range expansion (to regions where pathogens may be less established), alongside the warmth and drought that can be brought by climate change, can be expected to worsen spongy moth outbreaks by providing larger habitats and ideal conditions for population growth (Aouga & Simko-Bednarski, 2021).

A total of 33,000 km² have been treated with mating disruptors in the United States from 1993-2020 by the federal government’s Stop the Spread and suppression initiatives; another 25,000 km² has been treated with either Btk (19,000 km²), Dimilin (most common in the 20th century; 5,000 km²), or Gypcheck (500 km²; US Forest Service, 2022f, 2022g). Together, the state and federal governments have spent 283.4 million USD on suppression efforts since 1980, with spending by local governments and private owners unknown (US Forest Service, 2022h). The cost of suppression is estimated to be $8,600 per km² when applied (Tobin et al., 2012).

Eradication efforts (mediated via Btk application) may be more expensive, and generally occur only when populations are found far from the invasion front in new territories (likely the result of human-mediated dispersal). These usually are less than 4 km², but have ranged up to 1,800 km² in Lane County, Oregon in 1985 (with a cost of 9.5 million dollars for just that one program; Tobin et al., 2012). Overall, 29.8% of the area defoliated by spongy moths is treated in the US.⁷

Possible next steps

1. More accurately estimating the scale of spongy moth outbreaks. We don’t believe there’s anything about Pelham’s 2021 outbreak that should make it an outlier, and spongy moth extension and control experts we talked to agreed that 0.7-1.4 billion caterpillars per km² (3-6 million caterpillars/acre) are not unusual number for an outbreak. Using data collected by each state/province every year would provide better estimates.
   1. In addition, outbreaks do occur in the moth’s native range and presumably have very similar welfare issues. Future work could estimate the scale of outbreaks, and investigate the causes of suffering, in the native range.
2. Stopping the spread of non-native spongy moths into new areas is currently the best way to reduce future spongy moth outbreaks and suffering with the tools in hand. Mating disruptors are not likely to cause significant suffering and are generally effective at the low densities of the invasion front. The rate of spread is currently ~20%-50% of pre-1990s levels, and this program receives significant government funding. While some funding to support more effective/extensive monitoring programs or spongy moth control methods at the invasion front may be useful (such as improved mating disruption technology), it is not the most neglected or important area of spongy moth welfare.
3. Identifying interventions for preventing outbreaks as climate conditions shift, instead of simply suppressing outbreaks until natural population control returns, will be the most important task for improving spongy moth welfare long term. Currently, very little money or attention is dedicated to eradication of spongy moth outbreaks in areas behind the invasion front, as eradication is currently an expensive and challenging undertaking. Just a few research ideas in this area are:
   1. Research that reliably predicts when an outbreak will occur, so that overwintering eggs can be targeted for eradication (reducing the number of caterpillars suffering through peak and decline years in the outbreak).
   2. Technology that reliably kills eggs, preferably in a species-specific manner, and that can be applied to large areas of land at a relatively low cost (e.g., aerial).
   3. Technology that sterilizes adults or prevents mating in higher-density populations, reducing the overall population.
4. Identifying the highest-welfare solutions to outbreak suppression where outbreaks do occur. This is likely the most tractable and important research area for improving spongy moth welfare during outbreaks.
   
   1. Research could identify which of the currently-available suppression strategies is the best for insect welfare under a range of control scenarios.
   2. Research could focus on improving suppression technologies - for example, identifying significantly faster-acting suppression agents that can be applied as an aerial spray, replacing Btk or reliance on E. maimaiga and LdNPV. Shorter time-to-death may reduce suffering here, though the effectiveness and feasibility of these new methods (as well as non-target effects) would need to be assessed.
5. More research into the welfare impacts of spongy moth outbreaks on individuals of other species, both during the outbreak and in the years following the outbreak decline, would benefit our understanding of this complex ecological issue.
6. In general, the spongy moth is only one of many insect species around the world that can reach very high densities (wireworm, spotted lanternfly, locusts, bark beetles, termite colonies, etc.). Most anthropogenic and natural methods of bringing these insect populations back under control likely cause prolonged suffering. More research on insect population control methods, broadly, may be warranted in order to improve the welfare of many wild insects that regularly experience pest management protocols. However, it is important to note that research on insect population control is not neglected, per se. Rather, it is research on insect population control that directly considers the welfare impacts alongside efficacy and cost that is neglected.

Conclusion

This report demonstrates the importance of the spongy moth outbreak problem in North America, and the importance of wild insect welfare in outbreaks more generally. Trillions of insects are involved across thousands of km², generally experiencing negative welfare states for at least five to seven days, with outbreak cycles occurring every few years. Additionally, the problem is only getting larger - the invasion front is steadily increasing the geographic range of spongy moths, and climate change is reducing abiotic conditions that prevent outbreaks (very cold winter temperatures that kill eggs, and the cool/wet springs that promote E. maimaiga control of pre-outbreak level populations). Not only is the scale of the problem incredibly large in terms of the sheer numbers of individuals, but also the suffering these individuals experience is likely substantial whether undergoing anthropogenic or non-anthropogenic control.

Some aspects of L. dispar welfare may also be neglected. Annually, millions of dollars are dedicated to the suppression of smaller patches of L. dispar outbreaks in ecologically-important or human-dominated landscapes behind the invasion front. Additionally, millions of dollars are spent to stop the spread of the species to new areas, with limited success. Obviously, these are not neglected areas of research and intervention. However, to our knowledge, little funding is dedicated to researching more humane methods of outbreak suppression behind the invasion front (for the purposes of being more humane; e.g., faster-acting pesticides, sterilization technology, or control methods that work prior to egg hatch).⁸ In addition, there are currently fewer incentives to try and prevent an outbreak behind the invasion front before it occurs, rather than treat the most important, affected acreage after an outbreak happens. The technology to reliably and cheaply kill eggs does not yet exist, meaning that there’s less payoff for early detection of an upcoming outbreak event. In addition, the economic costs of large-scale prevention/eradication treatments would often be higher than the economic costs of allowing the spongy moth outbreak to naturally collapse in two to five years due to starvation and disease, given the large amount of commercially unused land (e.g., national park forests, private land, etc.) these outbreaks typically cover. Thus, 1) preventing outbreaks by targeting eggs prior to hatch in infested areas, and 2) researching more humane methods of anthropogenic control, are the most neglected aspects of the L. dispar outbreak welfare problem.

However, improving spongy moth welfare during outbreaks is unlikely to be immediately tractable, without investment in additional research. Ultimately, the best way to improve non-native L. dispar welfare during outbreaks is to eliminate the occurrence of outbreaks entirely - not just control the spread into new areas or quickly suppress ongoing outbreaks. Ecologists, forestry professionals, and government officials are all aligned with this goal, where the species is not native. Cost-effective, high-welfare solutions to reduce the occurrence of outbreaks long-term, or halt the creep of L. dispar into new areas, would thus be received positively by the broader community. However, given that there are significant incentives already in place to research ways to eradicate spongy moth outbreaks, it seems unlikely that additional funding would quickly and cost-effectively uncover eradication technologies. Additionally, even if eradication solutions were uncovered, the additional funding required to implement them on such a large geographic scale would likely also be tremendous, and therefore may not ever be tractable. The lack of tractability given current management tools is likely why outbreak prevention in areas where L. dispar is established is neglected, relative to stopping the spread and suppression efforts. It may be more tractable to instead focus on improving the welfare of caterpillars that are already under anthropogenic management (e.g., 29.8% of the total acreage defoliated by L. dispar in North America is treated). Currently, aerial Btk treatments are the most widely used technology for suppression - which kill larvae slowly via sepsis or starvation. Improving welfare in this percentage of the population could mean advocating for using or researching/developing faster-acting insecticides, insecticides that target eggs before they hatch, or developing better reproductive control technologies than the current mating disruptors that only work in low-density areas.

Overall, we hope this post demonstrates the potential scale of wild insect suffering, the various dimensions of its predictability, and the possibility that humans could play some role in mitigating it. On scale, this is just one insect species of many that undergo huge outbreaks that are affected by a range of likely-unpleasant and slow-acting mechanisms. On predictability, these outbreaks are cyclical: we understand their basic trajectories and can anticipate they will occur with some frequency. On mitigation, while there may not be any tractable interventions to take at this time, we know both that humans control spongy moths at significant rates and that there hasn't been any empirical research on spongy moth welfare. It is plausible, then, that relatively small investments in research could pay significant dividends. Regardless, bringing awareness to the scale and neglectedness of wild insect welfare is itself a neglected cause. We thus hope that we have contributed to bringing some new awareness to the problems faced by wild insects in outbreak conditions.

Caveats

1. This post is not meant to be exhaustively detailed (approximately 50 hours of work went into writing this post and the associated research). We do not go into great depth about many of the kinds of control methods used, the many different parasitoids, and their welfare impacts. We stick to the most commonly reported methods for brevity.
2. We assess only the number of caterpillars living in outbreak peak and decline conditions for this post; not the total number of spongy moth caterpillars alive in the wild, in total. The vast majority of spongy moth-occupied acreage has a lower-than-outbreak density of caterpillars every year. However, because suffering related to anthropogenic control, as well as suffering related to natural causes of death like LdNPV, increase significantly in outbreak conditions, we focus on outbreaks and not all caterpillars.
3. We scale up the Town of Pelham case study because few data sources provide as much population density detail as their 2020 report. In many cases, when an area is severely defoliated, the data provided are simply “2,000+ egg masses/acre (~0.5 million egg masses/km²)” in state or federal reports, indicating 2,000 egg masses/acre is considered by that state/province to be the minimum population density responsible for severe defoliation. However, as demonstrated by the Pelham report, the lowest population density that causes severe defoliation (e.g., ~0.5 million egg masses/km²) was not close to the average spongy moth population density at severely defoliated sites (~4.1 million egg masses/km²). In fact, the numbers in Pelham in 2019 go as high as 11.7 million egg masses/km² at some sites. Therefore, actual numbers of caterpillars may be much higher than can be assessed using state/federal agency data, because state agencies ‘stop counting’ at 2,000 egg masses/acre and may seriously underestimate the average. Additionally, different states/entities have different cutoffs for the minimum population density necessary to cause severe defoliation (which may relate to different forest characteristics, or simply different reporting guidelines), complicating quick inter-entity data analysis. Finally, we ignore the species’ native range entirely in this work.
4. Federal, state/province, and local governments, as well as private landowners, may all engage in control. Data on private landowners is especially challenging to find, and thus treatment costs and acreage are likely underestimated.
5. Much of our reference list comes from extension or federal/state government websites, which are not peer-reviewed scientific publications but are likely accurate for roughly understanding the scale of the spongy moth outbreak issue.

Credits

This report is a project of Rethink Priorities. The post was written by Meghan R. Barrett and Hannah McKay. Meghan was responsible for researching insect ecology and structuring the draft while Hannah was responsible for modeling outbreak numbers in Guesstimate and determining the prevalence of anthropogenic control. Thanks to Bob Fischer, Daniela R. Waldhorn, Jason Schukraft, Kim Cuddington, William McAuliffe, and the friendly extension experts and entomologists who shared their extensive expertise about spongy moth control and biology with us, for their various contributions.

If you are interested in our work, please consider subscribing to our newsletter. You can explore our completed public work here.

References

Andresen, J., Mccullough, D., Potter, B. E., Koller, N., Bauer, L., Lusch, D., & Ramm, C. W. (2001). Effects of winter temperatures on gypsy moth egg masses in the Great Lakes region of the United States. Agricultural and Forest Meteorology, 110, 85–100. https://doi.org/10.1016/S0168-1923(01)00282-9

Aouga, K., & Simko-Bednarski, E. (2021, July). Gypsy moths are stripping trees bare in the Northeast. Here’s why the outbreak is so bad. Retrieved September 15, 2022, from CNN website: https://perma.cc/5CWL-H88J

Arnaud, P. H. (1978). A host-parasite catalog of North American Tachinidae (Diptera). Department of Agriculture, Science and Education Administration. Retrieved from https://perma.cc/T5SC-3CH6

Chen, C.-J., Quentin, M. E., Brennan, L. A., Kukel, C., & Thiem, S. M. (1998). Lymantria dispar Nucleopolyhedrovirus hrf-1 Expands the Larval Host Range of Autographa californica Nucleopolyhedrovirus. Journal of Virology, 72(3), 2526–2531. https://doi.org/10.1128/jvi.72.3.2526-2531.1998

Collins, J. (n.d.). European Gypsy Moth. College of Agriculture, University of Kentucky. Retrieved from https://perma.cc/J2ZU-BNLM

Davidson, C. B., Gottschalk, K. W., & Johnson, J. E. (1999). Tree mortality following defoliation by the European gypsy moth (Lymantria dispar L.) in the United States: A review. Forest Science, 45(1), 74–84. https://doi.org/10.1093/forestscience/45.1.74

Davidson, C. B., Gottschalk, K. W., & Johnson, J. E. (2001). European gypsy moth (Lymantria dispar L.) outbreaks: A review of the literature. USDA Forest Service. https://perma.cc/N64B-HE5Z

Entomological Society of America. (2021). Entomological Society of America Discontinues Use of Gypsy Moth, Ant Names. Retrieved September 14, 2022, from https://perma.cc/B855-3V2B

Forbush, E., & Fernald, C. (1896). The gypsy moth. Porthetria dispar (Linn.). A report of the work of destroying the insect in the commonwealth of Massachusetts, together with an account of its history and habits both in Massachusetts and Europe. Boston: Wright & Potter Printing Co.

Frank, K. L., Tobin, P. C., Thistle, H. W., & Kalkstein, L. S. (2013). Interpretation of gypsy moth frontal advance using meteorology in a conditional algorithm. International Journal of Biometeorology, 57(3), 459–473. https://doi.org/10.1007/s00484-012-0572-4

Gibbons, M., Crump, A., Barrett, M., Sarlak, S., Birch, J., & Chittka, L. (2022). Chapter Three - Can insects feel pain? A review of the neural and behavioural evidence. In R. Jurenka (Ed.), Advances in Insect Physiology (Vol. 63, pp. 155–229). Academic Press. https://doi.org/10.1016/bs.aiip.2022.10.001

Gottschalk, K. W. (1990). Gypsy moth effects on mast production. In: McGee, Charles E., Ed. Proceedings of the Workshop: Southern Appalachian Mast Management; 1989 August 14-16; Knoxville, TN. Knoxville, TN: University of Tennessee: 42-50., 42–50. https://perma.cc/5E8F-M27Y

Grayson, K. L., & Johnson, D. M. (2018). Novel insights on population and range edge dynamics using an unparalleled spatiotemporal record of species invasion. Journal of Animal Ecology, 87(3), 581–593. https://doi.org/10.1111/1365-2656.12755

Howe, H. J. B. (2019). Improving Pest Management for Wild Insect Welfare. Wild Animal Initative. Retrieved April 17, 2023, from https://perma.cc/WT5S-4QZQ

Invasive Species Centre. (2022). Spongy Moth – Profile and Resource. Retrieved September 14, 2022, from Invasive Species Centre website: https://perma.cc/RJ9X-R6JN

Lallemand Inc./Bioforest. (2021). 2020 Gypsy Moth Monitoring Program, Town of Pelham, 2020 Population Assessments and 2021 Forecasts. Retrieved from https://perma.cc/AXY9-S3ZK

Lance, D. R., Elkinton, J. S., & Schwalbe, C. P. (1986). Feeding rhythms of gypsy moth larvae: Effect of food quality during outbreaks. Ecology, 67(6), 1650–1654. https://doi.org/10.2307/1939096

Mahr, S. (1999). Compsilura concinnata, parasitoid of gypsy moth. Midwest Biological Control News Online., 6(9).

McCullough, D. (2022a). Spongy Moth Life Cycle. Retrieved September 15, 2022, from MSU Extension, Integrated Pest Management website: https://perma.cc/58RG-QFFZ

McCullough, D. (2022b). A Virus and a Fungal Disease Cause Spongy Moth Outbreaks to Collapse. Retrieved September 15, 2022, from MSU Integrated Pest Management website: https://perma.cc/ZBB2-E4F4

McManus, M., Schneeberger, N., Reardon, R., & Mason, G. (n.d.). Gypsy Moth Forest Insect & Disease Leaflet 162. USDA Forest Service. Retrieved from USDA Forest Service website: https://perma.cc/ZZ46-MJ85

New Jersey Dept. of Agriculture. (n.d.). Gypsy Moth Suppression Q and A. Retrieved September 15, 2022, from https://perma.cc/74G2-CK8R

NYS Dept. of Environmental Conservation. (n.d.). Spongy Moth. Retrieved September 15, 2022, from https://perma.cc/2TYY-45EH

Ober, H., K. (2020). The Value of Oaks to Wildlife. Retrieved September 15, 2022, from Institute of Food and Agricultural Sciences Extension, University of Florida website: https://perma.cc/7QZZ-58XA

Pembroke Today. (2021, November). 146,000 ha of forests in Renfrew County defoliated by LDD moth in 2021. Retrieved September 15, 2022, from 104.9 Pembroke Today website: https://perma.cc/FTM2-NK6G

Robinson, D. (2021). Gypsy moth caterpillars are ravaging upstate NY trees, raining down feces: “It’s biblical.” Democrat & Chronicle. Retrieved from https://perma.cc/Y3DS-BZGE

Rowe, A. (2020). Insects raised for food and feed—Global scale, practices, and policy. Retrieved September 15, 2022, from Rethink Priorities website: https://rethinkpriorities.org/publications/insects-raised-for-food-and-feed

Sanford, J. L., Barski, S. A., Seen, C. M., Dickens, J. C., & Shields, V. D. C. (2014). Neurophysiological and Behavioral Responses of Gypsy Moth Larvae to Insect Repellents: DEET, IR3535, and Picaridin. PLoS ONE, 9(6), e99924. https://doi.org/10.1371/journal.pone.0099924

Smithsonian. (n.d.). Numbers of Insects (Species and Individuals). Retrieved September 15, 2022, from Smithsonian Institution website: https://perma.cc/7YDE-JSZV

Strazanac, J. S., Plaugher, C. D., Petrice, T. R., & Butler, L. (2001). New Tachinidae (Diptera) Host Records of Eastern North American Forest Canopy Lepidoptera: Baseline Data in a Bacillus thuriengiensis Variety kurstaki Nontarget Study. Journal of Economic Entomology, 94(5), 1128–1134. https://doi.org/10.1603/0022-0493-94.5.1128

The Connecticut Agricultural Experiment Station. (2021). Gypsy Moth 2021 Outbreak in Northwest Connecticut Causes Extensive Defoliation [Press Release]. Retrieved from https://perma.cc/KG7U-6WMW

Tobin, P. C., Bai, B. B., Eggen, D. A., & Leonard, D. S. (2012). The ecology, geopolitics, and economics of managing Lymantria dispar in the United States. International Journal of Pest Management, 58(3), 195–210. https://doi.org/10.1080/09670874.2011.647836

Toronto and Region Conservation Authority. (n.d.). LDD Moth. Retrieved September 14, 2022, from Toronto and Region Conservation Authority (TRCA) website: https://perma.cc/ZCG4-ZTS8

University of Illinois Extension. (n.d.). Gypsy Moth Management. Retrieved September 15, 2022, from University of Illinois Extension Gypsy Moth Northeastern Illinois Reporting Site website: https://perma.cc/8G4G-4LU8

University of Wisconsin-Madison. (n.d.). Spongy Moth (Lymantria dispar) in Wisconsin: Management guide for homeowners. Retrieved September 14, 2022, from Spongy Moth (Lymantria dispar) in Wisconsin website: https://perma.cc/AX2Y-MYCZ

US Forest Service. (2022a). Lymantria dispar dispar (Spongy Moth). Retrieved September 14, 2022, from USDA Forest Service Lymantria dispar dispar (Spongy moth) website: https://perma.cc/YC65-DFCJ

US Forest Service, U. (2022b). Spongy Moth Spread in North America. Retrieved September 14, 2022, from USDA Forest Service Spongy Moth: Spongy moth Spread in North America website: https://perma.cc/LDG8-4DL4

US Forest Service. (2022c). Lymantria dispar Digest. Retrieved September 15, 2022, from https://perma.cc/AF4K-S2UJ

US Forest Service. (2022d). Gypchek—The Spongy Moth NPV product. Retrieved September 15, 2022, from https://perma.cc/3KHZ-5ABK

US Forest Service. (2022e). Lymantria dispar Digest Defoliation Chart by Year. Retrieved September 15, 2022, from USDA Lymantria dispar Digest website: https://perma.cc/7RS8-2VTQ

US Forest Service. (2022f). Lymantria dispar Digest Slow The Spread by Year. Retrieved September 15, 2022, from https://perma.cc/T2PJ-42NH

US Forest Service. (2022g). Lymantria dispar Digest Suppression by Year. Retrieved September 15, 2022, from USDA Lymantria dispar Digest website: https://perma.cc/KTJ7-ND9P

US Forest Service. (2022h). Lymantria dispar Digest Suppression Cost by Year. Retrieved September 15, 2022, from https://perma.cc/C6RT-R6D4

Webb, R. E., White, G. B., Sukontarak, T., Podgwaite, J. D., Schumacher, D., & Reardon, R. C. (2004). Biological efficacy of Gypchek against a low-density leading-edge gypsy moth population. Northern Journal of Applied Forestry. 21(3): 144-149., 21(3), 144–149.

Yamasaki, M. (2003). White pine as wildlife habitat. Managing White Pine in a New Millennium 2003 Workshop Proceedings; 2003 October 9-10; Hillsborough, NH. Durham, NH: University of New Hampshire Cooperative Extension: 33-36. Hillsborough, NH. Durham, NH: University of New Hampshire Cooperative Extension: https://perma.cc/2U5K-QNYD

Notes

---

1. Moths that die during outbreak years two and three may have worse lives than those that live and die outside of outbreak conditions, as quick deaths like predation likely represent a lower proportion of the population. Btk deaths are almost exclusively related to outbreak or spread management, and cause prolonged suffering. A higher proportion of the moths in an outbreak will die slowly of LdNPV (which is density-dependent) and starvation. Parasitism is likely to be high in both cases. Finally, even in years where the outbreak is cresting, the lives of caterpillars in a high competition outbreak environment can be expected to be very stressful, inducing significant behavioral and physiological shifts such as changes in feeding habits (time or host plant; Lance et al., 1986) or increased susceptibility to disease.  ↩

2. Mating disruptors are chemicals that can prevent male insects from finding mates (thus reducing the species’ reproductive success). In spongy moths, mating disruptors are chemicals that mimic female pheromones; these chemicals are broadly applied to the environment and make it harder for males to locate females using these cues. Mating disruptors are only successful in areas where population density is low, and thus, males need to rely on chemosensory cues to find mates.  ↩

3. The surveyed sites may not represent the unsurveyed areas perfectly. However, the majority of unsurveyed acreage is likely to be rural/forested; the surveyed sites that best match these characteristics all had heavy/severe defoliation. Therefore, we expect our approach (which weights the surveyed town sites more heavily than their actual representation in the overall acreage of Pelham) is likely to underestimate, rather than overestimate, the number of caterpillars in the overall area.  ↩

4. These are the lower confidence limit of the lower estimation and the higher confidence limit of the higher estimation, respectively ↩

5. We assume that the 5,864 km² of moderately to severely defoliated area of Ontario represents 74% of the outbreak area in 2020, as this was the case for Pelham in that year (the rest of the area having ‘nil’ or ‘light’ defoliation). This makes the outbreak area 7,924 km² for 2020. Similarly, we assume that 17,000 km² is 75% of the outbreak area in 2021. Therefore, we take the outbreak area for 2021 to be 22,666 km², to account for light defoliation in our estimates. ↩

6. Although Pelham was in a ‘decline’ year in 2021, there is no reason to assume that the rest of Ontario is also in a ‘decline’ year (especially given that moderate-to-severe defoliation acreage increased more than threefold in the rest of Ontario from 2020 to 2021, we might expect that some of the rest of Ontario is in ‘peak’ years and not yet declining). However, to be conservative in our calculations, we use the profile of the ‘decline’ year from Pelham.  ↩

7. Based on USFS data for their Eradication, Stop the Spread and Suppression programs, see spreadsheet here.  ↩

8. Most commercial insecticides, for example, those applied for agricultural pest control, cause suffering, and their application is an even larger-scale issue than non-native spongy moth outbreaks each year (Howe, 2019). However, working on solving L. dispar welfare during control efforts is likely more tractable than working on insect welfare in relation to the use of commercial insecticides broadly. 1) In agricultural settings, there is a diversity of insect species targeted. Solving the problem would require a diversity of new, species- or taxa-specific humane insecticides and potentially even multiple new application methods. The spongy moth is a single species for which we already have a large amount of information on its biology, feeding behavior, etc. to inform delivery methods. This makes developing a more humane insecticide for just this one species much more tractable than solving the broad-action commercial insecticide problem. In addition, people care deeply about what is sprayed on/near their food, while many fewer individuals care about what is sprayed on their uninhabited forest lands. There is likely to be less resistance (and potentially even many allies) to new, humane spongy moth control or eradication solutions, as compared to new agricultural pest control methods. Finally, many categories of agricultural pesticides could already be considered more humane than the viral/fungal diseases, Btk sprays, and starvation used to bring spongy moth outbreaks under control, as they are often much faster-acting (though this is not the case when Bt GMOs are used agriculturally). ↩