Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
Rats and the COVID-19 pandemic: Early data on the global emergence of rats in
response to social distancing
Michael H. Parsons1, Yasushi Kiyokawa2, Jonathan L. Richardson3, Rafal Stryjek4¸ Kaylee A.
Byers5,6, Chelsea G. Himsworth6,7, Robert M. Corrigan8, Michael A Deutsch9, Masato Ootaki2,
Tsutomu Tanikawa10, Faith E. Parsons11,12, Jason Munshi-South13
1Department of Biological Sciences, Fordham University, Bronx, NY, USA
2Laboratory of Veterinary Ethology, The University of Tokyo, Tokyo, Japan
3Department of Biology, University of Richmond, Richmond, VA, USA
4Institute of Psychology, Polish Academy of Sciences, Warsaw, Poland
5Department of Interdisciplinary Studies, University of British Columbia, Vancouver, Canada
6Canadian Wildlife Health Cooperative, The Animal Health Centre, British Columbia, Canada
7British Columbia Ministry of Agriculture, Abbotsford, Canada
8RMC Pest Management Consulting, Briarcliff Manor, New York, USA
9Medical and Applied Entomology, Arrow Exterminating Company, Inc. Lynbrook, NY, USA
10Tokyo Pest Control Association, Tokyo, Japan
11CareSet Systems, Houston, TX, USA
12Center for Behavioral and Cardiovascular Health, Columbia University, New York, New York,
USA
13Department of Biological Sciences and the Louis Calder Center—Biological Field Station,
Fordham University, Armonk, NY, USA
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Abstract
Following widespread closures of food-related businesses due to efforts to curtail the spread of
SARS-CoV-2, public health authorities reported increased sightings of rats in close vicinity of
people. Because rats vector a number of pathogens transmissible to people, changes in their
behavior has consequences for human health risks. To determine the extent of how stay-at-home
measures influenced patterns of rat sightings we: 1) examined the number of rat-related public
service requests before and during the period of lockdown in New York City (NYC) and Tokyo,
Japan; 2) examined reports made in proximity to closed food service establishments in NYC; and
3) surveyed pest control companies in the United States, Canada, Japan, and Poland. During the
month following lockdown, the overall number of reports decreased by 30% in NYC, while
increasing 24% in Tokyo. However, new hotspots of 311 calls were observed in proximity of
closed food service establishments in NYC; and there was a consistent positive association
between kernel density estimates of food service establishments and location of 311 calls (r =
0.33 to 0.45). Similarly, more reports were observed in the restaurant-dense eastern side of
Tokyo. Changes in clientele for pest control companies varied geographically, with 37% of pest-
management companies surveyed in North America reporting 50-100% of their post-lockdown
rat-related requests coming from new clients. In Warsaw, where there are no clusters of
restaurants in densely-populated areas, there were no changes. In Tokyo, there were no changes
in clients. We conclude that changes in public service calls are region-specific and localized,
with increases in rat sightings more likely near restaurant-dense regions. Pest control companies
surveyed in North America either lost much of their business or shifted clientele from old to new
locations. We discuss possible mitigation measures including ramping up pest control during re-
opening of food-related establishments and the need for citywide rodent surveillance and disease
monitoring.
Keywords: COVID-19, disease ecology, global commensal rodents, pandemic, public health,
rodent emergence, rodent surveillance, urban hygiene
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Introduction
Urban rats (Rattus spp.) are global commensal organisms that depend on humans for food and
harborage. Thus, shifts in human behavior, such as occurs following natural catastrophe or
pandemics, will have pronounced effects on nearby rat populations. This effect has been
hypothesized after social distancing and business closures were broadly enacted in Spring 2020
to limit the spread of SARS-CoV-2. Shortly thereafter, local governments and public health
authorities around the world began reporting that closures of restaurants and food-related venues
have coincided with reports of mass sightings of rats1-3. These sightings include reports of
aggressive behaviors during daylight hours and in close proximity to people4,5 with some rats
consuming conspecifics (e.g. muricide or cannibalism;6. Given the heightened state of anxiety
among people affected by the social distancing mandates7,8, these sightings could also be a
reflection of human sensitivity toward rats as indicators of poor sanitation and disease9,10.
Further, due to the increased use of social media during the pandemic11,12 and the rate at which
‘panic’ spreads as a social contagion globally13, predisposition or cognitive bias could lead to
increased reports. These reports, while broadly circulated in the popular media, have yet to be
reviewed or confirmed through research or surveillance.
The potential mass movements of rats into new areas may negatively impact human
society14,15. Rats are known to transmit many types of disease14,16,17 and are associated with
billions of dollars in losses of food annually18,19.They are commonly thought to have killed more
humans than all wars combined20,21, and these fearful perceptions have caused an intense fear of
rats in many cities22. The mere presence of rats is enough to cause harm to mental well-being,
particularly in low socio-economic status areas10,23. But despite these harmful consequences,
there are no validated methods to quantify rat movements irrespective of human reports24-26.
Further, there are no routine disease surveillance programs to detect changes in pathogens that
rats may be transmitting14. Thus, cities are limited to the reporting of rat sightings to determine
spatial risks of rodent-borne disease and where abatement programs should be concentrated27,28.
While cities need new mechanisms to monitor rat populations, it is important to
understand the relationship between human behavior and rat activity. Changes in the activity of
either species may have profound implications for the other22,23,27. In temperate regions where
seasonal changes are prominent, rats and humans concurrently increase activity during Spring
and Summer, which results in predictable increases in rat sightings. For instance, reports of rat
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
complaints in Chicago over a ten-year period, 2008-2018, demonstrated a particularly strong rise
from winter to spring, then peaks each summer and decreases again from fall to winter29. These
seasonal fluctuations in complaints can be explained by both decreased activity of rats as they
retreat deep within their burrows in colder months30,31 and the reduction in human outdoor
activity in winter32,33. Human reporting of rat sightings can also be impacted by contextual bias
and cognitive bias34. For instance, rats observed in new environments are more likely to be
noticed and reported than rats in areas where they have previously been established. Further, a
single rat sighting in a public area might generate more complaints compared to those in less
visible, or private, areas15. Lastly, the bandwagon effect35 is a type of predisposition that causes
people who have knowledge of a widely-communicated event such as the global media reports
on rats, to have an increased awareness of rats, and thus, be more likely to report them.
Because humans have undergone profound changes resulting from social distancing and
isolation and because commensal rat populations are commensal with humans, rat populations
are almost certainly affected, however the nuances of these impacts remain unknown. Our
intentions for this paper are to provide data to help us understand to what extent changes in
human reporting behavior and/or wild rat ecology were responsible for the widely-reported
phenomenon. In colder climates, such as found in New York City, we sought to identify
observable shifts in public service reports (e.g., 311 calls) independent of seasonal effects29.
Calls to 311 are free to report. Calls reported to pest management companies36 on the other hand,
convey an added sense of urgency as they carry a monetary expense. Due to detailed, publicly
available data for the city, we were also able to examine food availability for rats in the form of
restaurants and food carts. In Tokyo we utilized public calls to the Tokyo Pest Control
Association (TPCA) which are used as a first step prior to hiring an exterminator, along with
cost-carrying reports made directly to private pest control companies. Neither Poland nor Canada
have immediately accessible public service reports, but national pest management associations
exist in each country, which we used to distribute surveys.
Methods
We analyzed public service requests (311 calls) from NYC Open Data
(https://opendata.cityofnewyork.us/) on May 3, 2020 for the observation period beginning
January 1, 2014 and ending April 30, 2020, inclusive of the lockdown phase which began on
March 23, 2020. New York City was considered an appropriate region for analysis due to its
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
high human population density, and robust, regularly-updated and easily accessible reporting
data. Reports are made when people file a complaint through the city’s 311 via phone, website or
smartphone app. The selected interval was chosen because it allows a before-and-after treatment
analysis to be considered alongside annual seasonal changes. Rodent-related 311 calls are
classified into 5 categories: signs of rodents, conditions attracting rodents, rat sighting, mouse
sighting, and rodent bites. For this paper, we limited our analysis to the “rat sighting” category.
In Tokyo, the Japanese Government requested self-isolation on Feb. 16 and then declared
a state of emergency on Apr. 7. Therefore, we analyzed the number of phone calls to the TPCA
from January through April from 2015 through 2020, which was provided by the TPCA. The
TPCA is a non-profit public corporation established by the Ministry of Health & Welfare of the
Japanese Government. As a public interest incorporated association, they provide services and
referrals to the general public and also have access to 105 pest-management companies.
Complaints to the TPCA and subsequent consultation are free of charge, other than cost of a
local call. The TPCA introduces its members to pest control operations when necessary.
However, people are still responsible to decide whether or not they will contract with a pest-
management companies.
Pest control interventions offer a means to determine prevalence of rats36. Thus, we also
created an industry survey with the help of industry professionals (Appendix A). The survey
covered at least a 30-day period from lockdown to post-lockdown for each nation. Three primary
questions were posed: 1) whether there had been changes in overall rat-related calls from
customers; 2) what approximate proportion of post-lockdown customers were new customers
(which may account for customers in new residences not previously infested); and 3) whether the
number of new customers for rat jobs was different from the number of new customers during
the same period the previous year. Respondents were not required to disclose personally
identifiable information. The survey was distributed through pest control channels (CPCA)
throughout Canada on May 5, 2020 and throughout the United States via Pest Control Magazine
4 as well as via social media through Twitter and LinkedIn. The survey was translated in
Japanese and distributed to the 105 TPCA member pest-management companies on May 11,
2020. Because Japan has a large number of roof rats, the survey additionally asked about the
proportion of roof rats, brown rats, and house mice being reported. The survey was also
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
translated in Polish for dissemination via pest management channels, but due to the low numbers
of roof rats in Poland, there was no request for the species of rat being reported.
Statistics
Phone calls
We used one-way ANOVA to identify overall differences in average daily rat sightings per
month across all years in the observation period, and Tukey’s HSD test for pairwise comparison
between months. We then used two-way ANOVA to test the association between the daily
number of rat sightings in the months of March and April during the previous 6 years without
social distancing, 2014 to 2019, and during the social distancing phase in year 2020. We used
Tukey’s HSD test to conduct pairwise comparisons between the periods without social
distancing and the period with social distancing for the month of March, and similarly for the
month of April.
The Tokyo metropolitan area includes 23 wards, 26 cities, 3 towns, and 1 village. Among
them, the “accommodation, eating, and drinking services” (79.9% in 2016, based on Tokyo
Metropolitan Government website) and population (69.2% in 2020, based on Tokyo
Metropolitan Government website) were heavily biased toward the eastern 23 wards area (Fig.
2). Therefore, we focused on the data in 23 wards. From 2018 to 2020, the number of calls in the
target month was compared with the average of previous 3 years in the same month. When the
number of calls in 2020 was increased or decreased by more than 2 standard deviation as
compared to the previous 3 years, the number was classified as “increased” or “decreased”,
respectively. The differences in the number of wards showing “increased”, “decreased,” and “no
change” among 2018, 2019, and 2020 were analyzed by Fisher’s exact test.
Spatial analysis
To investigate changes in the spatial distribution of rats in NYC over the study period, we
mapped 311 rat calls in GIS and applied hot spot analysis. Both ArcMap 10.6 and QGIS 3.10
were used for these analyses. We first divided 311 calls for NYC by month and year and
imported these as separate data layers in GIS. We generated a time series of 311 rat calls in
QGIS by creating a heatmap and working within the Time Manager plugin. We then conducted
an optimized hot spot analysis in ArcMap using the Mapping Clusters toolbox. This analysis
evaluates areas with more or less point occurrences than expected at random, and applies a Getis-
Ord Gi statistic to assess significant deviations from a random distribution. We used the NYC
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
borough outline as the bounding polygon for this analysis, and a fishnet pattern was used to
generate sampling polygons across the study area, as recommended for this tool. Lastly, we
created kernel density raster layers of both the food service establishments (using NYC Open
Data on inspections) and rat 311 calls by month, and used the Band Collection Statistics toolbox
to calculate correlation coefficients between each raster dataset.
Survey
Descriptive statistics were used to report survey findings. In Tokyo, a two-way repeated
ANOVA was used to analyze the proportion of rodents. Planned comparisons were conducted
using paired t test to clarify the changes caused by a state of emergency and Tukey HSD test to
clarify the predominant species before and after a state of emergency.
Results
311 calls
In New York City, one-way ANOVA showed a significant difference in daily rat sightings
among the 12 months throughout the observation period (F=106.5; P<0.001). Tukey’s HSD tests
indicated that, from 2014 to 2018, there are five groups of months with similar number of rat
sightings that appear to follow a seasonal trend. Specifically, the lowest number of daily rat
sightings were lowest in the winter burrowing periods (December and January) while the highest
were in summer (July through August). April sightings were significantly different from all the
other months, while March and November were similar (Fig 1A).
Two-way ANOVA showed a significant association between the average number of rat
sightings for the months of March and April during the years without social distancing (2014 to
2019) compared to the current year (2020) where social distancing was enforced (Fig 1B,
F1=35.4; P<0.001). For the month of March, the early part of the COVID-19 lockdown period
where social distancing was enforced, there was no significant difference in the average number
of sightings during the social distancing phase ( 24.9; SD=4.4) as compared to March in the
prior 6 years when there was no social distancing ( 35.6; SD=13.6; P=0.1). However, for the
month of April, the first full month of isolation orders in NYC, the number of rat sightings
during the social distancing period ( 31.7; SD=8.7) was significantly lower compared to
April in the prior 6 years without social distancing ( 45.0; SD=14.8; P<0.0001). Although
the number of rat sightings in April 2020 was significantly lower compared to the same month in
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
prior years, the figure shows an increasing trend in the reported rat sightings from March to April
2020 (consistent with the trend shown in Fig. 1A) however, the difference is not statistically
significant (P=0.560).
Spatial relationship with food service establishments
There was a positive correlation between the kernel density estimates of food service
establishments and location of 311 calls in all of the months assessed (Fig. 4; Supplemental
Table 1). However, the spatial association between food service establishments and rat 311
locations did increase between the pre- and post-lockdown periods (r of 0.35–0.36 in
January/February 2020 compared to 0.44–0.45 in March/April 2020). Additionally, the Getis-
Ord analysis identified hotspots of 311 complaints within NYC, which changed only modestly
before and after the COVID lockdown period (Fig. 4). The correlation analysis also showed a
high correlation among 311 calls throughout the 3 years (r = 0.83 to 0.95), which indicates that
the locations of the rodent complaints are generally consistent across the years and months
analyzed. There were no detectable hotspots in areas away from clusters of food service
establishments (Supplemental Fig. 1).
Calls to TPCA (Tokyo)
Within the 23 wards of interest in Tokyo, the total number of calls from January through April.
were 224 (2015), 296 (2016), 293 (2017), 299 (2018), 314 (2019), and 365 (2020) which ranged
from 47.3% to 54.2% of calls in the Tokyo Metropolis. In the 23 wards, the total number of
phone calls was increased in Jan. and Apr. 2020, but not changed in February and March. The
patterns of changes in Tokyo were shown in Fig 2. In January, 5 of 6 changes (4 increased and 1
decreased), 2 of 4 changes (2 increased), and 1 of 4 changes (1 decreased) occurred in the 23
wards in 2018, 2019, and 2020, respectively. Fisher’s exact test revealed that the changes in
Tokyo (P = 0.55) and in 23 wards (P = 0.16) were not different among three years. In February,
5 of 5 changes (3 increased and 2 decreased), 5 of 7 changes (5 increased), and 6 of 10 changes
(5 increased and 1 decreased) occurred in the 23 wards in 2018, 2019, and 2020, respectively.
Fisher’s exact test revealed that the changes in Tokyo (P = 0.23) and in 23 wards (P = 0.72) were
not different among three years. In March, 3 of 5 changes (2 increased and 1 decreased), 2 of 4
changes (2 decreased), and 4 of 7 changes (4 increased) occurred in the 23 wards in 2018, 2019,
and 2020, respectively. Fisher’s exact test revealed that the changes in Tokyo (P = 0.08), but not
in 23 wards (P = 0.16), were significantly different among three years. In April, no change
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
occurred in 2018. While 1 of 2 changes occurred in the 23 wards in 2019, 5 of 7 changes
occurred in 2020 (5 increased). Fisher’s exact test revealed that the changes in Tokyo (P < 0.05)
and in 23 wards (P < 0.05) were different among three years.
Survey Data
There were 50 respondents from North America, 26 from Poland, 85 from Tokyo, and one each
from Argentina, India, and Malaysia. Of the 50 respondents in North America, one responded
“prefer not to answer” to the first question on whether there has been changes in number of rat-
related calls from customers, 3 to the second question on whether the majority of post-lockdown
customers were new customers, and 5 to the third question on whether the number of new
customers for rat jobs was different to the number of new customers during the same period the
previous year. The responses from Argentina, India and Malaysia were excluded from the
analysis, and responses of “prefer not to answer” were excluded on a per question basis. Data
from pest management companies from the US and Canada were combined as there was a
similar pattern in responses. The 50 respondents were from 47 cities across North America.
Given the number of cities represented, compared to the total number of respondents, we assume
that we received one response per pest management company.
Of 85 respondents in Tokyo, 82 responses were included as all questions were answered
clearly. The 82 respondents covered of 29 of 53 areas of Tokyo and all 23 wards. We focused on
the 75 respondents from the 23 wards, because these wards, along with Tokyo, function as a
single urban entity with respect to public services such as rat complaints. The number of
questions answered ranged from 1 to 9 in each area. All respondents from Poland answered all
three questions and were all included in the analysis.
Volume of rat-related calls or jobs.
More than half of the respondents in North America (55%, Fig. 3A) indicated that they received
an increase in the volume of rat-related calls or jobs, while only 9% of responses from the 23
wards reported an increase. In Tokyo, the majority of the responses from the 23 wards (61%)
indicated that there was no change in the volume of calls. And in Poland, the distribution was
very similar between those who experienced an increase, a decrease and no change in the overall
volume of rat-related calls or jobs.
Relative volume of rat-related calls or jobs from new clients.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
More than half of the respondents from North America (Fig. 3B, 53%) reported that they have
more calls from new clients for rat-related jobs as compared to the previous year. This is in
contrast with the responses from Poland (8%) and Tokyo (7%) where the majority of the
responses indicated that the relative volume was the same as compared to the same time in the
previous year (54% and 71%, respectively).
Proportion of rat-related jobs from new clients.
When asked how many of the rat-related calls or jobs were from new clients, 13% of respondents
in North America indicated that all of their rat-related jobs were for new clients, while none of
the respondents from Tokyo or Poland selected this response (Fig. 3C). More than half of the
respondents in Poland (54%) answered that none of their rat-related jobs were from new clients,
while only 21% in Tokyo and 13% in North America selected the same answer. Most of the
respondents from Tokyo (60%) indicated that the proportion of rat-related jobs from new clients
were very low, 1% to 25%, while the same response was obtained from one-third of the
respondents from both Poland (35%) and North America (32%).
Among the 24 respondents in North America who experienced an increase in rat-related
calls or jobs, 21 (88%) indicated that they have acquired more new clients for rat-related jobs
compared to the same time last year, while the remaining 3 (12%) responded that the relative
volume is about the same.
Proportion of rodents in Tokyo
Before declaration of a state of emergency, the proportion of brown rats, roof rats, and house
mice in Tokyo 23 wards was 20.5 ± 1.7 %, 78.9 ± 1.8 %, and 0.7 ± 0.3 %, respectively. The
proportion changed to 21.6 ± 2.0 %, 77.8 ± 2.1 %, 0.7 ± 0.3 %, respectively, after the
declaration. The proportion was significantly affected by the type of rodents (F(2,225) = 748, P
< 0.01), but not by the declaration of a state of emergency (F(1,225) = 0, P = 1). Interaction of
these two factors was not significant (F(2,225)=1.32, P = 0.27).
A planned comparison by paired t test did not find significant changes by the declaration
of a state of emergency (brown rats: t(75) = 0.94, P = 0.35; roof rats: t(75) = -0.94, P = 0.35;
house mice: impossible to assess because the majority was 0). A planned comparison by Tukey’s
HSD test revealed that the proportion of roof rats was significantly higher than those of brown
rats (P < 0.01) and house mouse (P < 0.01) both before and after a declaration of a state of
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
emergency. The proportion of brown rats was higher than those of house mouse both before and
after a declaration of a state of emergency (P < 0.01).
Discussion
There are currently no validated measures to identify rat behaviors or movements in the city
despite their importance to human health risks. However, two key pieces of evidence suggest that
changes in human reporting behavior have to some extent been precipitated by changes in rat
movements. First, we have found a surprisingly high association between the location of rodent
complaints and food service establishments and; secondly, we provide evidence that these ‘hot
spots’ of rat sightings have shifted during the pandemic. Had these changes in rat reports been a
result of human cognitive bias34 and/or hypervigilance due to ‘the bandwagon effect’35, then
changes would have likely been unrelated to food businesses and more geographically uniform.
Instead, changes in public service calls and requests for pest control were region-specific and
localized, with increases in rat sightings in New York City (NYC) and Tokyo more likely to be
reported within, or near, restaurant-dense regions. In Warsaw, where there are no clusters of
restaurants within densely-populated areas, there were no changes.
The fact that calls decreased in NYC was unanticipated because institutions that reported
rat sightings1-3 implied there were many more sightings than usual. The decrease in sightings was
further surprising because the period of lockdown coincided with the spring season when rat
sightings typically increase as rats emerge from underground burrows29. Yet, the post-pandemic
reports moved in the opposite direction from the previous 6 years, with a 30% decrease in
volume of 311 calls in NYC. Conversely, in Tokyo, we were surprised that while calls actually
did increase by 23% as compared to 5 previous years, there were no differences between surveys
administered before and after the lockdown period. However, within North America, the shift in
clientele was striking. More than one-third of pest management companies surveyed reported 50-
100% of their post-lockdown jobs were from new clients, while six companies reported that
100% of their client-base had changed from old to new clients. Within the pest management
industry, such pronounced changes in rat populations following catastrophes are expected37,38.
Some possible explanations for the geographic variability in outcomes relate to
differences in rat control, human social behavior and the ecological differences between the
primary species of rats being reported. In NYC, the lockdown was accompanied by the
widespread adoption of social distancing practice. The highest concentrations of sightings
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
typically occur near subway lines and recreational public spaces39. However, in April 2020, the
majority of residents were at home social distancing and many had left the city (e.g., counter-
urbanization37). Additionally, due to the abrupt reduction in available food, many rat populations
had temporarily decreased due to stress and competition. Another possible explanation is that
people who found an infestation near their residences may have contacted pest control directly
rather than reporting to 311. This finding makes sense because prior to restaurant closures a
single rat might be seen by hundreds of patrons and reported. During business closures the
opposite might be expected to occur, where many rats in a single residence might elicit only a
single call—likely to pest control for immediate response and not 311.
In Tokyo, it was surprising that the increased phone calls to TPCA were not reflected in
increased business for pest-management companies. This might be because the two species of
rats present in Tokyo were affected differently by the lockdown as roof rats were responsible for
78% of complaints registered in the surveys, while Norway rats made up only 20%. In addition
to being adapted for warmer climates, roof rats are more difficult to spot; they are smaller,
arboreal, make nests inside of ceilings or roofs, and migrate shorter distances. Pest control in
Tokyo is also mandatory in large buildings and thus roof rats are most likely found inside smaller
buildings such as restaurants and bars. Because restaurants and bars remained open for ‘take-
away’ meals, roof rats could continue to feed on food stocks, grease vats, and oil stains. Calls
regarding this species would have not been impacted greatly by the lockdown.
In contrast, the eastern side of Tokyo, where calls increased, is composed of the downtown
area where restaurants are more concentrated. In this area, large amounts of garbage are placed
street-side at midnight where Norway rats can consume garbage ad libitum until collected the
following morning. Thus, it is highly probable that roaming by Norway rats increased the overall
number of phone calls to TCPA, but was not perceived as important enough to hire a private pest
control specialist. We should also note that in Tokyo most pest-management companies contract
with restaurants and bars in one-year installments. The businesses that utilize these contracts for
the majority of their clientele would not be expected to be affected much by the declaration.
Specifically, if the number of one-year contracts had been much larger than new contracts, then
surveys of pest control companies would have failed to reveal the changes caused by the
declaration. In any case, we found that pest-management companies were not impacted as
prominently as NYC. Poland does not have a public service request system; however, they do
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
enforce mandatory pest control with fines for offenders with rats on their properties. Given that
they do not have clusters of restaurant-dense areas in heavily-populated areas, it is reasonable that
there were no changes in pest control needs before or after lockdown.
If large numbers of rats are migrating in some areas or regions—as appears to be
supported by our findings—then certain characteristics of immigrant rat populations may
change40. For instance, if formerly insular rat populations41 come into increased contact with one
another due to migration, then mating opportunities could occur between colonies that do not
typically interbreed. Additionally, if large numbers of rats concurrently migrate, then a greater
variety of individual genotypes within populations would have moved significant distances while
under intense selective pressure40. It is reasonable then to assume there is some influence on the
genetic profile of future urban rat populations42,43. While stressed rat populations may initially
decrease due to loss of food and increased competition and muricide, rats breed rapidly, and
should rapidly re-establish population equilibrium44. Impacts on human populations are less
clear.
Rodents are considered unlikely candidates to be infected with or transmit SARS-CoV-
245. Only two rodent species, the atypical 5 kg giant bamboo rat (Rhizomys sinensis)46, and the
golden Syrian hamster (Mesocricetus auratus)47 have been associated with the COVID-19 virus..
However, stressed or wounded rodents are less likely to groom and keep their fur clean, and thus
present a greater hazard to mechanically transmit contagious diseases without being infected
themselves48,49. Additionally, in the case where rodents transmit pathogens that affect the
immune system, such as Lyme disease, then those affected could be at increased risk of a
weakened immune response to other pathogens50,51. These increased risks to human health
support the urgent need for more robust data on rats in cities so that we can adequately track
these changes and make informed decisions moving forward regarding management.
Conclusions
Reports of increased sightings of urban rats near food service establishments suggest that
lockdown procedures have resulted in increased rat movement, which could, in turn, alter the
distribution of rats throughout the city. Because rats carry a number of pathogens that are
transmissible to people, these changes may have potential negative impacts for humans.
However, it is difficult to determine the extent to which social-distancing measures have
impacted rat biology due to a general lack of systematic data collected on rat presence and
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
abundance. To overcome these limitations, we combined 311 data, geographic data, and survey
data to infer whether rat sightings have increased in relation to stay-at-home measures put in
place during the COVID-19 pandemic. We demonstrate that following lock-down measures in
NYC, that rat complaints decrease as compared to previous years. Further we show that changes
in calls to pest control companies in North America, Japan, and Poland vary. Interestingly and
importantly, we have demonstrated that that in North America there is a striking increase in the
number of calls from new clientele, suggesting that stay-at-home measures have resulted in
noticeable changes in rat presence.
Natural disasters typically result in temporary decreases in rat populations followed
shortly by an uptick in populations37,38. Residents near food service establishments should
therefore be prepared for local increases in rat populations during the late spring and early
summer. Properties closest to restaurants should pay special attention to exclusion techniques
and pest control to ward off the increased risks. General mitigation measures include public
education regarding urban hygiene and widespread, systemic changes in how we manage our
garbage52. We recommend that all commensal species53 be tested for their ability to harbor or
transmit the virus, especially until a vaccine is developed. These assays could be particularly
useful because rats frequent sewers where they come into contact with human wastes which were
recently shown to contain heavy loads of the COVID-1954,55. Given the impact of rats on public
health, and the challenges of exploiting human reports to learn about rats, there is a clear need
for more consistent, rigorous, and standardized rat monitoring and disease-monitoring on a
global scale.
Acknowledgements
We thank the Cornell University Extension Service (Matt Frye), Canadian Pest Management
Association (Amy Cannon), Pest Control Technology (Brad Harbison), and National Pest
Association Fairfax, Virginia (Jim Fredericks) for support and internal distribution of surveys.
Prof. Stanislaw Ignatowicz helped distribute surveys in Poland.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Funding Statement
This work was entirely self-funded by the authors.
References
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Figures
Figure 1. Right, the association between number of reported rat sightings per day by month with reference to the social distancing
period of March and April 2014 to 2019 compared with the same two months during the social distancing period, year 2020. Left,
distribution of daily rat sightings per month during the observation period (January 2014 to April 20) using data from NYC Open
Data.
20)
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Figure 2. New and old clients. Survey responses from 53 pest control companies in North America, 26 in Poland, and 75
in 23 wards
of Tokyo, Japan. Top right, number of all rat-related calls/jobs from new clients. Top left, relative volume of overall rat-related
calls/jobs before and after the lockdown period. Bottom, proportion of new client calls due to rat-related jobs.
ds
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Figure 3. The changes in service calls to Tokyo Pest Control Association over period from
January through April from 2018 to 2020. The grey areas indicate the 23 wards.
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Figure 4. Kernel density of food service establishments (top left) and results of optimized
hotspot analysis (Getis-Ord Gi*) of 311 calls in New York City, USA (right and bottom).
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
Supplemental Table 1. Correlation matrix between the food service establishment (FSE) kernel density raster and monthly 311 calls
kernel density raster in New York City, USA.
FSE Mar 2018
Apri 2018 Mar 2019
Apr 2019 Jan 2020
Feb 2020 Mar 2020
Apr 2020
FSE 1
Mar 2018 0.40 1
Apri 2018 0.33 0.95
1
Mar 2019 0.42 0.88
0.89 1
Apr 2019 0.44 0.90
0.92 0.93
1
Jan 2020 0.36 0.90
0.92 0.91
0.92 1
Feb 2020 0.35 0.83
0.87 0.88
0.91 0.91
1
Mar 2020 0.44 0.88
0.91 0.90
0.92 0.92
0.90 1
Apr 2020 0.45 0.84
0.89 0.90
0.90 0.91
0.89 0.95
1
. CC-BY-NC-ND 4.0 International licenseIt is made available under a
is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted July 7, 2020. .https://doi.org/10.1101/2020.07.05.20146779doi: medRxiv preprint
