The role of deforestation and fragmentation in yellow fever virus dispersal

doi:10.21203/rs.3.rs-1956824/v1

Download PDF

Research Article

The role of deforestation and fragmentation in yellow fever virus dispersal

https://doi.org/10.21203/rs.3.rs-1956824/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Context The intense process of deforestation in tropical forests poses serious challenges for the survival of biodiversity, as well as for the human species itself. This scenario is supported by the increase in the incidence of epidemics of zoonotic origin observed over the last few decades. In the specific case of sylvatic yellow fever (YF), it has already been shown that an increase in the transmission risk of the causative agent (yellow fever virus - YFV) is associated with areas with a high degree of forest fragmentation, which can facilitate the spread of the virus.

Methods In this study we tested the hypothesis that areas with more fragmented landscapes and a higher edge density (ED) but a high degree of connectivity between forest patches favor YFV spread. To this end, we used YF epizootics in non-human primates (NHPs) in the state of São Paulo to build direct networks, and used a multi-selection approach to analyze which landscape features could facilitate YFV spread.

Results Our results showed that municipalities with the potential to spread the virus exhibited a higher amount of forest edge. Additionally, the models with greater empirical support showed a strong association between forest edge density and the risk of occurrence of epizootic diseases, as well as the need for a minimum threshold of native vegetation cover to restrict their transmission.

Conclusions These findings corroborate our hypothesis that more fragmented landscapes with a higher degree of connectivity favor the spread of YFV, while landscapes with fewer connections tend to act as dead zones for the circulation of the virus.

Sylvatic yellow fever

Haemagogus leucocelaenus

Forest edge

Edge density

Landscape metrics

Tropical forests are home to most of the planet’s biodiversity (Reed et al., 2020). The Amazon Forest, for example, accounts for around one quarter of all the species in the world (Betts et al., 2008), and the Atlantic Forest is on the list of biodiversity hotspots, areas that have a high concentration of endemic species and are suffering intense habitat loss (Myers et al., 2000). These areas are also responsible for providing what are known as ecosystem services, a wide range of benefits that mankind obtains from nature and which are fundamental for its well-being (MEA, 2005), such as climate regulation, the provision of water, pollination and disease regulation (Charles and Dukes, 2007).

Deforestation has direct impacts not only on biodiversity loss, but also on the provision of ecosystem services (Brockerhoff et al., 2017; Crespin and Simonetti, 2016; Mori, 2017), leading to a reduction in the quality of life and health status in human populations as a result of increased exposure to and contact with pathogens (Plowright et al., 2017), particularly those of zoonotic origin.

Zoonotic diseases are infectious diseases transmitted from animals to humans through exposure to zoonotic pathogen by direct contact, food, water or the environment (WHO, 2021) and correspond to 60% of existing infectious diseases (Jones et al., 2008). The increase in the number of epidemics of zoonotic origin over the last decades and the speed with which the COVID-19 epidemic evolved to become a pandemic indicate that regulating services provided by ecosystems are not working correctly and that man-made changes to the landscape, especially deforestation and changes in land use, are the ultimate causes of the increase in the number of these epidemics (Gottdenker et al., 2014; Jones et al., 2013; Mishra et al., 2021). Indeed, around 40% of recent zoonotic diseases were the result of deforestation and the associated changes in land use (Loh et al., 2015).

A number of studies have attempted to shed light on how vegetation cover and changes in land use affect the risk of transmission of zoonotic diseases. For example, deforestation across Southeast Asia has been suggested to be the main factor responsible for dispersal of Plasmodium knowlesi, one of the causative agents of malaria (Davidson et al. 2019). Changes in land cover together with climate changes have affected the abundance of the vectors and hosts of Japanese encephalitis virus (Le Flohic et al., 2013), and man-made changes to the landscape have led to increased abundance of mosquito vectors and changes in the transmission patterns of West Nile virus (Ferraguti et al. 2021). However, few studies have investigated this issue in the Atlantic Forest (Ogrzewalska et al., 2011; Prist et al., 2021a; Prist et al., 2017, 2016; Scinachi et al., 2017), mainly because of the complexity of the host-pathogen dynamics involved in zoonoses and landscape changes. Studies in this field involving YF and its causative agent are even scarcer. Worthy of note are the studies by Abreu et al. (2022), Childs et al. (2019), de Almeida et al. (2019), Ilacqua et al. (2021) and Prist et al. (2021b), of which the latter two identified the risk of transmission of YFV associated with thresholds of vegetation cover of 30–70% and forest edge density (ED) > 80 m/ha, and greater mobility of the virus associated with forest edges, respectively. In a recent study of Atlantic Forest patches, our research group found greater permeability to YFV in forest edges than in core areas classified as Water, Human-impacted, Urban and Forest (Wilk-da-Silva et al., 2022), corroborating earlier results reported by Prist et al. (2021b).

Changes in land use are a key driver of habitat reduction, with a consequent negative impact on biodiversity (Fahrig, 2003) and direct influence on the dilution effect. This effect predicts that the greater the species diversity, the less the chance of contact between host and vectors of a particular type of pathogen. A reduction in species diversity can therefore lead to increased abundance of host and reservoir populations, increasing the probability of host-vector interaction and consequently the likelihood of infection (Ostfeld and Keesing, 2000; Schmidt. and Ostfeld, 2001).

In the particular case of zoonoses caused by mosquito-borne pathogens, their distributions involve the presence of hosts, vectors and human populations (Dantas-Torres and Otranto, 2015). Hence, overlapping of these plays a determining role in spillover events, an excellent example being sylvatic YF, a disease that affects humans when they are bitten by infected mosquitoes after entering forest areas (Monath and Vasconcelos, 2015). The disease is caused by YFV, an arbovirus in genus Flavivirus (Flaviviridae) which circulates in the wild among non-human primates (NHPs) and mosquitoes in the family Culicidae, in particular genera Haemagogus and Sabethes in Central and Southern America and Aedes (subgenus Diceromyia and Stegomyia) on the African continent (Chippaux and Chippaux, 2018; Monath and Vasconcelos, 2015).

Deforestation also affects the risk of transmission of this disease as loss of forest cover usually leads to fragmentation of natural environments, resulting in greater forest edges (Fahrig, 2003). The edges have a warmer, dryer microclimate and favor habitat generalist species, such as one of the main vectors of YF, Haemagogus leucocelaenus (Dyar and Shannon, 1924) (Abreu et al., 2019a), which exhibits ecological plasticity to vegetation patches subjected to man-made disturbances (Gomes et al., 2010; Lopes, 1997; Silva and Lozovei, 1999), as well as other potential vector species (Couto-Lima et al., 2017; Wilk-da-Silva et al., 2020). Furthermore, the juxtaposition of remnant forest patches and modified areas can contribute to greater availability of sources of blood meals for these vectors (Mucci et al., 2015), leading to increased abundances. Edge environments also make it easier for humans to enter vegetation patches (Bloomfield et al., 2020) and consequently have more contact with animals and any pathogens these may carry.

As well as contributing to an increase in edge surface, habitat fragmentation is directly reflected in the incidence and abundance patterns of vector mosquito species (Zittra et al., 2017) and causes changes in interaction patterns between hosts, vectors and pathogens (Zohdy et al., 2019). In the specific case of YF, habitat fragmentation is a key factor in the maintenance and circulation of YFV, particularly because of a greater prevalence of the virus due to the increased density of NHP populations (Mbora and McPeek, 2009).

The Atlantic Forest, with its highly fragmented remnants (Rezende et al., 2018), therefore consists of areas in which YFV can easily spillover to human populations as these areas are home to hosts (NHPs) and vectors (mosquitoes of genus Haemagogus and Sabethes) of this virus (Abreu et al., 2022; Possas et al., 2018), with both sometimes overlapping in edge areas (Bolt et al., 2018; Lenz et al., 2014; Wilk-da-Silva et al., 2020). This scenario is corroborated by the virus’s capacity to persist in Atlantic Forest remnants during consecutive transmission intervals (Abreu et al., 2019b), together with the large number of human cases associated with these remnants, which are close to cities with high population densities.

Zoonoses are generally one of the main public-health problems in low-income countries, where they account for approximately one quarter of all infectious diseases, and have less impact in higher-income countries, as a result of which they tend to be neglected (Grace et al., 2012). An understanding of aspects of landscape involved in the maintenance and circulation of pathogens such as YFV is crucial to support epidemiologic surveillance services in order to ensure better allocation of human and material resources in the fight against epidemics.

The main aim of the present study was therefore to understand how landscape structure affected the dispersal of YFV in the state of São Paulo between 2016 and 2020 using an interaction network analysis approach. Our hypothesis is that municipalities that act as a source of the virus and are therefore responsible for disseminating it to other locations in the state have a more fragmented landscape with greater forest edge density and less isolated patches than municipalities where the virus is not disseminated. This occurs because a more fragmented yet more connected landscape can favor dissemination of the virus, whereas fragmented landscapes with fewer connections do not allow the virus to spread, instead acting as a dead zone and preventing YF outbreaks in the state.

2.1. Study area

The state of São Paulo extends over 248,219.48 km² (24.8 million ha) and has an estimated human population of 46,649,132 and a population density of 166.25 inhabitants/km² (IBGE, 2021). Currently, the state consists predominantly of agricultural areas – 17,778,910 hectares (ha) (71.6%); native vegetation – 5,577,114 ha (22.5%), of which 3.9 million ha consist of Atlantic Forest and 163,674 ha of Cerrado; areas without vegetation – 880,768 ha (3.54%); and water bodies – 582,619 ha (2.34%) (MapBiomas, 2020; SOSMA, 2020).

2.2. Data on epizootics

Between 2016 and 2020, the state of São Paulo confirmed a total of 713 epizootics of YF in NHPs. Yellow fever is a disease for which immediate notification is compulsory, i.e., every suspect event, be it the death of a NHP or a human case with symptomatology compatible with the disease, should be reported within 24 hours to the relevant local authorities, whose responsibility it is to notify the Ministry of Health. Data at municipal level were obtained from the Ministry of Health center for general coordination of epidemiologic surveillance under the law of access to information, ref. no. 25820004039202025. For our analyses we used only those events for which there was a date of occurrence (n = 702) and ignored those for which no date was given (n = 11). The first record was in the northwest region of the state, in the municipality of Potirendaba (07/10/2016), while the last event was reported in Greater São Paulo, in the municipality of Barueri (10/03/2020) (Fig. 1, Supplementary Table 1).

2.3. Landscape data

To extract the landscape metrics, we used the Brazilian Foundation for Sustainable Development (BFSD) land-use map (https://www.fbds.org.br/). The BFSD base map has a total of six land-use classes: 1) Forest Formation - Tree vegetation native to the Atlantic Forest with a continuous canopy; 2) Water - Continuous water surface; 3) Human-impacted Area - Areas without any native vegetation cover; 4) Built Area - Areas with buildings (https://www.ibge.gov.br/geociencias/cartas-e-mapas/bases-cartograficas-continuas/15759-brasil.html), 5) Non-forest Formation - Native bushy or herbaceous vegetation and 6) Forestry - Eucalyptus sp. or Pinus sp. plantations. The map was created by means of supervised classification based on RapidEye images from 2013 (scale 1:10,000). The metrics extracted, which are based on forest-cover classes, were Percentage of the landscape occupied by the forest area (PLAND); Forest Patch Density (PD); Forest Edge Density (ED), which was determined considering a distance of 150 m from the edge of the forest patch toward its interior (Wilk-da-Silva et al., 2022); Largest Forest Patch Index (LPI); Percentage of Like Adjacencies (PLADJ); and Euclidean Nearest Neighbor Distance, ENN), both for forest areas. These metrics were collected for each of the municipalities for which there was any record of an epizootic. The metrics PLADJ and ENN were calculated based on an estimate of the dispersal radius expected by Jackson and Fahrig (2012) of 0.3 to 0.5 times the maximum dispersal distance of a species. The radius used therefore corresponded to 2,850 meters, half the maximum dispersal distance of the species Hg. leucocelaenus (Causey et al., 1950), the main vector in the state.

Next, buffers extending 11.5 km outwards around the centroids of each of the municipalities with occurrences of epizootics were used to capture the neighboring municipalities without any records of events during the same period and identify differences in landscape composition patterns compared with municipalities with records of epizootics. The distance chosen was based on the mean rate of dispersion observed for mosquitoes of the genus Haemagogus (Causey et al., 1950) as a way of sampling a representative number of municipalities without occurrences. The analyses were carried out with ArcGIS 10.5 and Fragstats 4.0.

2.4. Network analysis

Network analysis finds widespread use in science, including sociology, psychology, anthropology, biology and medicine, as it provides a conceptual model that allows connection patterns (links) to be identified from the nodes making up a particular network (Ames et al., 2011; Martínez-López et al., 2009). By tracing these interactions, a series of biological questions can be addressed. In parasitology, for example, the tool has been used to answer ecological and evolutionary questions involving host-pathogen interactions (Runghen et al., 2021). In epidemiology, connections between individuals or groups can be traced to gain an understanding of epidemiologic dynamics in the spread of infectious diseases (Danon et al., 2011) with a view to providing improved support for strategies to combat these. An example is the data generated from network analyses used in studies of the spread of Covid-19 (Loyal and Chen, 2020).

In order to estimate which municipalities could act as hubs for YFV in the state of São Paulo, networks were constructed in which the centroids of the municipalities were considered as nodes and the differences in days between epizootic events as links. Using this definition, we chose to include in our analyses only links between municipalities separated by distances of 30 km or less and with a period of 30 days or less between records of epizootics. These figures were based on an estimated mean daily rate of spread for YFV of 1 km (Hill et al., 2020).

For a municipality to be considered a hub, the connection value had to be greater than the mean (384.1), while non-hubs had values lower than the mean. Once the hubs had been identified, they were grouped together and their landscape structures were compared with those of municipalities with fewer connections (non-hubs).

2.5. Statistical analysis

The data were tested for normality with the Shapiro-Wilk normality test, and Mann-Whitney non-parametric tests were applied to assess the differences in landscape metrics observed in the municipalities, which were separated into hub and non-hub groups and groups with and without records of epizootics. Binomial “logit” generalized linear models using a binary response variable (occurrence or non-occurrence of epizootics) and predictor variables (landscape metrics associated with forest patches) were constructed from municipalities with occurrences of epizootics (n = 76) and municipalities without occurrences (n = 86) as a way of determining the extent to which these variables contributed to the occurrence of epizootics. The continuous variable PLAND was transformed into a dichotomous variable using the minimum threshold values of vegetation cover associated with the risk of transmission of YF (30–70% - high risk; <30% or > 70% – low risk), as described in Ilacqua et al. (2021). In addition, the inflation factor for each predictor variable and possible collinearities between variables were observed; when these had high values (> 0.4) the variables were removed from the analysis. The information-theory approach based on the Akaike information criterion (AIC) was applied to select the most plausible statistical models. The models with the smallest AIC were considered the best, and ∆AIC ≤ 2 was adopted as the cutoff to select models with more empirical support. The strength of evidence in favor of each model was evaluated using Akaike weights (Burnham and Anderson, 2002).

Network analysis was performed with Gephi 0.9.2 (https://gephi.org), and statistical analysis was carried out in the R version 4.1 computing environment (R Development Core Team, 2019). The following packages were used: mass (Venables and Ripley, 2002); bbmle (Burnham and Anderson, 2002); car (Fox and Weisberg, 2018); and ggplot2 (Wickham, 2016).

Our results show that between 2016 and 2020 there were epizootics in 76 (11.8%) municipalities out of a total of 645 municipalities in the state of São Paulo. The municipality with the greatest number of epizootics was Mairiporã (n = 95), followed by São Paulo (n = 76), Jundiaí (n = 60), Itapecerica da Serra (n = 49) and Bragança Paulista (n = 47). Based on the parameters used in the study, a network with 76 nodes and 244 links was built in which each municipality was adjacent to on average three others (3.2 nodes), i.e., the virus spreads from one municipality to approximately three others. There were on average 384.1 links for each municipality, with Mairiporã having the largest number of connections (15,415), followed by the municipalities of São Paulo (8,509); Jundiaí (5,758); Atibaia (4,976) and Itapecerica da Serra (4,804) (Supplementary Table 1, Supplementary Fig. 1).

The results for landscape metrics based on the degree of forest cover were as follows. PLAND for the municipalities with the most links varied between 27% and 54%: Mairiporã – 53%; São Paulo – 27%; Jundiaí – 36%; Atibaia – 32%; and Itapecerica da Serra – 54%. Patch density (number of patches per 100 ha, PD) varied as follows: Mairiporã 2.14; São Paulo 1.61; Jundiaí 2.63; Atibaia 4.16; and Itapecerica 2.81 (Supplementary Fig. 2).

The figures for ED were: Mairiporã − 74.9 m/ha; São Paulo − 34.5 m/ha; Jundiaí − 39.9 m/ha; Atibaia − 59.3 m/ha; and Itapecerica da Serra − 77.3 m/ha. The proportion occupied by the largest forest fragment in the landscape (LPI), where values close to 0 correspond to highly fragmented landscapes, and values close to 100 to landscapes formed by a single patch, varied from 5–33% in these municipalities: Mairiporã − 20.5%; São Paulo − 15.5%; Jundiaí − 24.1%; Atibaia − 5.96%; and Itapecerica da Serra − 33.1%. The proportion of like adjacencies (PLADJ) varied between 85.74% and 91.43%; the figure for Mairiporã was 89.15%; for São Paulo, 89.75%; Jundiaí, 91.43%; Atibaia, 85.74%; and Itapecerica da Serra, 88.91%. The range between the minimum and maximum Euclidean nearest neighbor (ENN) distances for the forest fragments closest to these municipalities varied from 750.0 to 1,988.6 m: the figure for Mairiporã was 1,088.2; for São Paulo, 1,988.6; Jundiaí, 1,434.6; Atibaia, 750.0; and Itapecerica da Serra, 224.6 (Supplementary Fig. 2).

When the metrics for municipalities with a larger number of connections (hubs) were compared with those for municipalities with fewer connections (non-hubs), only ED exhibited a statistically significant difference [w = 763; p-value = 0.008023; 95% confidence interval (CI): 4.111612 − 23.589589; difference in location (dl) = 14.9365] (Table 1, Fig. 2C). This indicates that municipalities with a large number of connections and therefore greater potential for virus dispersal have on average 15 m of ED per ha more than municipalities with fewer connections (Table 1, Fig. 2C).

When the landscape metrics for municipalities with records of epizootics and those without epizootics were compared, once again there was a statistically significant difference in ED [w = 4254; p-value = 0.0009414; 95% confidence interval (CI): 4.486643 − 18.348575; difference in location (dl) = 11.95587], as there was for the variables PLAND [w = 4274; p-value = 0.0007392; 95% confidence interval (CI): 4.067215 − 14.397460; difference in location (dl) = 9.287526]; LPI [w = 4085; p-value = 0.006138; 95% confidence interval (CI): 0.3107364 − 4.6418871; difference in location (dl) = 1.393859] and PLADJ [w = 4015; p-value = 0.01223; 95% confidence interval (CI): 0.5072591 − 4.9484638; difference in location (dl) = 2.717798] (Table 2, Fig. 3).

Comparison of multiple models based on landscape metrics for municipalities with and without occurrences of epizootics showed that at least six models were equally plausible (model07, model09, model02, model6, model04 and model08). Model07 had the greatest empirical support and strength of evidence (ΔAIC = 0, Weight = 0.2348) and an additive effect between the variables ED and LPI. The former exhibited a strong association with the occurrence of epizootics (slope = 0.5436; p-value < 0.01), while the latter also had a positive association but with a weaker individual contribution (slope = 0.2909; p-value = 0.07603) (Table 3, Fig. 4). It should be noted that the variable ED is also present in the models with the greatest strength of evidence, model09, model02 and model06.

Table 1

Comparison (non-parametric Mann-Whitney test) of landscape metrics associated with the degree of forest cover in municipalities with records of epizootics in the state of São Paulo divided into those from which the virus was spread (hubs) and those from which it was not (non-hubs). A period of no more than 30 days between records of YF epizootics in NHPs and a maximum distance of 30 km between the centroid of the municipalities were used (statistically significant values are shown in bold, p < 0.05).
Class	w-value	p-value	95% CI	dl
PLAND	629	0.2967	-5.300177 − 15.043149	5.82504
PD	580	0.6485	-0.6245647 − 1.0084316	0.18065
ED	763	0.008023	4.111612 − 23.589589	14.9365
LPI	618	0.3619	-3.414654 − 6.849423	1.1352
PLADJ	575	0.6922	-2.956582 − 4.745092	0.84052
ENN	560.5	0.8244	-181.6325 − 232.3993	14.77273
Percent of Landscape – PLAND (%); Patch Density – PD (number per 100 ha); Edge Density – ED (m/ha) and LPI - Largest Patch Index (percentage of the landscape corresponding to the largest patch); Percentage of Like Adjacencies - PLADJ (number of like adjacencies involving Forest core areas); and Euclidean Nearest Neighbor Distance – ENN (distance to the nearest neighboring patch). CI - confidence interval; dl - difference in location.

Table 2

Comparison (non-parametric Mann-Whitney test) of landscape metrics associated with the degree of forest cover in municipalities with records of epizootics in the state of São Paulo and neighboring municipalities (based on maximum displacement of mosquito species of genus *Haemagogus*) without any records of epizootics (statistically significant values are shown in bold, p < 0.05).
Class	w-value	p-value	95% CI	dl
PLAND	4274	0.0007392	4.067215–14.397460	9.287526
PD	3689	0.1582	-0.1583909 − 0.8786594	0.3645658
ED	4254	0.0009414	4.486643–18.348575	11.95587
LPI	4085	0.006138	0.3107364–4.6418871	1.393859
PLADJ	4015	0.01223	0.5072591–4.9484638	2.717798
ENN	2738	0.07555	-247.24615–12.64109	-113.4631
Percent of Landscape – PLAND (%); Patch Density – PD (number per 100 ha); Edge Density – ED (m/ha) and LPI – Largest Patch Index (percentage of the landscape corresponding to the largest patch); Percentage of Like Adjacencies – PLADJ (number of like adjacencies involving Forest core areas); and Euclidean Nearest Neighbor Distance – ENN (distance to the nearest neighboring patch). CI – confidence interval; dl – difference in location.

Table 3

Models tested and their parameters listed according to the best fit for occurrences of epizootics in the state of São Paulo. The models are based on landscape metrics associated with forest cover in 162 municipalities (76 with occurrences of epizootics and 86 without). The standard error of the estimate is shown in parentheses.
Model		Intercept	Slope	z value	p-value	AIC	ΔAIC	df	weight
Model07						215.0	0	3	0.2348
Response variable	Explanatory variable
Epizootic		-0.1326 (0.1651)		-0.803	0.42178
	Edge Density		0.5436 (0.1717)	3.166	0.00155 **
	Largest Patch Index		0.2909 (0.1640)	1.774	0.07603
Model09						215.2	0.2	3	0.2101
Epizootic		-0.1420 (0.1653)		-0.859	0.39031
	Edge Density		0.4859 (0.1760)	2.761	0.00576 **
	Percentage of Like Adjacencies		0.3131 (0.1851)	1.692	0.09071
Model02						216.2	1.2	2	0.1289
Epizootic		-0.1293 (0.1634)		-0.791	0.428706
Epizootic	Edge Density		0.5612 (0.1694)	3.312	0.000925 ***
Model06						216.4	1.4	4	0.1168
Epizootic		-0.1427 (0.1665)		-0.857	0.391
	Largest Patch Index		0.2331 (0.1814)	1.285	0.199
	Edge Density		0.4054 (0.2505)	1.618	0.106
	Euclidean Nearest Neighbor Distance		-0.2334 (0.3225)	-0.724	0.469
Model04						216.9	1.9	3	0.0920
Epizootic		-0.2516 (0.1894)		-1.328	0.184
	Percentage of Landscape		0.3565 (0.4183)	0.852	0.394
	Euclidean Nearest Neighbor Distance		-0.6378 (0.2596)	-2.457	0.014 *
Model08						216.9	1.9	3	0.0896
Epizootic		-0.1644 (0.1663)		-0.989	0.32274
	Largest Patch Index		0.1465 (0.1785)	0.821	0.41172
	Euclidean Nearest Neighbor Distance		-0.6581 (0.2526)	-2.605	0.00917 **
Model05						217.6	2.6	3	0.0639
Epizootic		-0.173044 (0.166045)		-1.042	0.29734
	Patch Density		-0.006154 (0.170294)	-0.036	0.97117
	Euclidean Nearest Neighbor Distance		-0.737426 (0.247522)	-2.979	0.00289 **
Model10						218.6	3.6	3	0.0388
Epizootic		-0.1562 (0.1644)		-0.95	0.34214
	Patch Density		0.3877 (0.1857)	2.088	0.03680 *
	Percentage of Like Adjacencies		0.6219 (0.2120)	2.933	0.00335 **
Model03						220.3	5.3	4	0.0164
Epizootic		-0.2761 (0.1891)		-1.460	0.1442
	Percentage of Landscape		0.5906 (0.3912)	1.510	0.1311
	Patch Density		0.3671 (0.1906)	1.926	0.0541
	Largest Patch Index		0.4487 (0.1930)	2.324	0.0201 *
Model0						222.8	7.8	2	0.0048
Epizootic		-0.3309 (0.1836)		-1.802	0.0715
Epizootic	Percentage of Landscape		0.8417 (0.3746)	2.247	0.0247 *
Model01						223.7	8.7	2	0.0030
Epizootic		-0.1213 (0.1596)		-0.76	0.447
Epizootic	Largest Patch Index		0.3344 (0.1668)	2.005	0.045
Null Model						226.0	11	1	< 0.001
Epizootic		-0.1236 (0.1574)		-0.785	0.432
df = Degrees of freedom; * is significant at p < 0.001 level; is significant at p < 0.01 level,* is significant at p < 0.05 level.

Municipalities with records of epizootics and more connections (hubs), and therefore a greater chance of receiving and spreading YFV, had an ED per ha approximately 15 m higher than the corresponding figure for municipalities with fewer connections (non-hubs). This pattern was also observed when municipalities with records of epizootics were compared with neighboring municipalities without a record; in this case, the municipalities with occurrences had an ED per ha on average approximately 12 m greater than the corresponding figure for municipalities without occurrences. When the models for predicting the risk of epizootics were compared, the variable ED was again found to be a key factor and was observed in all the more plausible models together with the variables PLADJ, ENN and LPI. This corroborates our hypothesis that more fragmented landscapes—with consequently higher ED—whose patches are closer to each other have a greater tendency to spread YFV. In addition, a minimum vegetation threshold (PLAND) was found to be necessary for there to be a risk of epizootics occurring, as observed by Ilacqua et al. (2021) in a recent study.

Recent studies in Atlantic Forest remnants have highlighted the role played by the edges of forest fragments in favoring YFV mobility (Prist et al., 2021b; Wilk-da-Silva et al., 2022) and in increasing the risk of transmission (Ilacqua et al., 2021). These findings indicate the importance of edge surfaces in YFV maintenance and transmission and suggest that these surfaces may be the main element in this equation. Forest-edge environments not only play an important role in the maintenance of populations of Hg. leucocelaenus, one of the main YF vectors in Southeast Brazil (Vasconcelos, 2010) because of this mosquito’s ecological plasticity to human-impacted forest environments (Gomes et al., 2010; Lopes, 1997; Silva and Lozovei, 1999), but also favor other potential vector species, as observed in the different patterns of abundance and occurrence of these species (Li et al., 2022; Wilk-da-Silva et al., 2020). Howler monkeys, for example, one of the main YF hosts in South America (Bicca-Marques et al., 2020), circulate in these environments (Bolt et al., 2021, 2019). Therefore, this type of environment can be highly favorable to maintenance of YFV. Furthermore, a larger edge surface can mean a greater area in which human populations can come into contact with the virus (Plowright et al., 2017).

The two main hubs (Mairiporã and São Paulo) are neighboring municipalities and home to Cantareira State Park (CSP), the largest urban forest in the world, which formerly had one of the largest populations of howler monkeys (Alouatta guariba clamitans) as well as Callicebus nigrifons, Sapajus nigritus and Callithrix aurita (Trevelin et al., 2007). The park, part of which is in the north of the city of São Paulo, is considered an important entry route for YFV into the city (Cunha et al., 2019; Mucci et al., 2016), which has a population of more than 12 million people (IBGE, 2021).

These two main hubs are followed by the municipalities of Jundiaí, Atibaia and Itapecerica da Serra, which all lie within the city of São Paulo Green Belt Biosphere Reserve (GBBR), a buffer zone for the São Paulo metropolis (Instituto Florestal, 2021). Paradoxically, the GBBR may end up being a source from which the virus can spread to the metropolis as human-generated pressures often found in vegetation patches in the middle of a built-up area (Figueiredo, 2017) can lead to changes in richness and abundance patterns for potential vector species (Wilk-da-Silva et al., 2020), as well as to the presence of Hg. leucocelaenus, albeit sometimes in lower abundance (Medeiros-Sousa et al., 2017). Additionally, the juxtaposition of patch edges and adjacent highways can affect the mobility of YFV, making it easier for the virus to spread (Prist et al., 2021b).

The results of the present study highlight the importance of preventing the advance of forest fragmentation in order to ensure landscape integrity and avoid spillover of pathogens of zoonotic origin into human populations (Plowright et al., 2021). In parallel, preserving existing forests could contain the spread of YFV, as observed by Wilk-da-Silva et al. (2022).

It should be pointed out that, unlike in the last study carried out by our research group, in which we used geographic coordinates obtained in the field in the locations where epizootics occurred (Wilk-da-Silva et al., 2022), in the present study we used approximations. This approach was intentional and was intended to ensure that the resulting methodology could be used in historical analyses or in longer historical series when the precise geographic locations of human cases and epizootics cannot be retrieved.

Fortunately, since 2019, when the virus arrived in the states of Southern Brazil, the geographic coordinates, as well as other information associated with the epizootics, have been recorded automatically in the field with the SIS-GEO/FIOCRUZ platform (MS, 2021). This initiative has undoubtedly improved the quality and speed of surveillance and studies involving not only YF, but also other zoonoses, thereby making real-time, spatially accurate data on these events available to the public.

Our results suggest that more fragmented landscapes with greater connectivity between forest patches increase the likelihood of YF epizootics and that the risk of these epizootics is dependent on a minimum vegetation threshold (PLAND). When municipalities with greater connectivity (hubs) were compared with those with less connectivity (non-hubs) and when those with records of epizootics were compared with those without, a pattern of greater ED per ha was observed in hubs as well as in municipalities with occurrences of epizootics.

Funding

This work was supported by São Paulo State Research Foundation (FAPESP, Grant numbers: 2018/25437–6, 2018/18751–6 and 2017/23127–7) and the Brazilian National Council for Scientific and Technological Development (CNPq, Grant number: 301865/2018–3).

Competing Interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Author Contributions

Ramon Wilk da Silva: Conceptualization, Methodology, Investigation, Writing - original draft, Formal analysis. Paula Ribeiro Prist: Conceptualization, Methodology, Investigation, Formal analysis, Writing - review & editing, Supervision. Antônio Ralph Medeiros-Sousa: Investigation, Formal analysis, Review & editing. Gabriel Zorello Laporta: Investigation, Data curation, Review & editing. Luis Filipe Mucci: Investigation, Visualization, Review & editing. Mauro Toledo Marrelli: Project administration, Funding acquisition, Writing - review & editing.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Abreu, F.V.S. de, de Andreazzi, C.S., Neves, M.S.A.S., Meneguete, P.S., Ribeiro, M.S., Dias, C.M.G., et al., 2022. Ecological and environmental factors affecting transmission of sylvatic yellow fever in the 2017–2019 outbreak in the Atlantic Forest, Brazil. Parasit. Vectors 15, 23. https://doi.org/10.1186/s13071-021-05143-0
Abreu, F.V.S. de, Ribeiro, I.P., Ferreira-de-Brito, A., Santos, A.A.C. dos, Miranda, R.M. de, Bonelly, I. de S., et al., 2019a. Haemagogus leucocelaenus and Haemagogus janthinomys are the primary vectors in the major yellow fever outbreak in Brazil, 2016–2018. Emerg. Microbes Infect. 8, 218–231. https://doi.org/10.1080/22221751.2019.1568180
Abreu, F.V.S. de, Delatorre, E., dos Santos, A.A.C., Ferreira-de-Brito, A., de Castro, M.G., Ribeiro, I.P., et al., 2019b. Combination of surveillance tools reveals that Yellow Fever virus can remain in the same Atlantic Forest area at least for three transmission seasons. Mem. Inst. Oswaldo Cruz 114. https://doi.org/10.1590/0074-02760190076
Ames, G.M., George, D.B., Hampson, C.P., Kanarek, A.R., McBee, C.D., Lockwood, D.R., et al., 2011. Using network properties to predict disease dynamics on human contact networks. Proc. R. Soc. B Biol. Sci. 278, 3544–3550. https://doi.org/10.1098/rspb.2011.0290
Betts, R.A., Malhi, Y., Roberts, J.T., 2008. The future of the Amazon: new perspectives from climate, ecosystem and social sciences. Philos. Trans. R. Soc. B Biol. Sci. 363, 1729–1735. https://doi.org/10.1098/rstb.2008.0011
Bloomfield, L.S.P., McIntosh, T.L. & Lambin, E., 2020. Habitat fragmentation, livelihood behaviors, and contact between people and nonhuman primates in Africa. Landsc. Ecol. 0123456789. https://doi.org/https://doi.org/10.1007/s10980-020-00995-w
Bolt, L.M., Russell, D.G., Coggeshall, E.M.C., Jacobson, Z.S., Merrigan-Johnson, C., Schreier, A.L., 2019. Howling by the river: howler monkey (Alouatta palliata) communication in an anthropogenically-altered riparian forest in Costa Rica. Behaviour 157, 77–100. https://doi.org/10.1163/1568539X-00003582
Bolt, L.M., Russell, D.G., Schreier, A.L., 2021. Anthropogenic edges impact howler monkey (Alouatta palliata) feeding behaviour in a Costa Rican rainforest. Primates 62, 647–657. https://doi.org/10.1007/s10329-021-00904-y
Bolt, L.M., Schreier, A.L., Voss, K.A., Sheehan, E.A., Barrickman, N.L., Pryor, N.P., et al., 2018. The influence of anthropogenic edge effects on primate populations and their habitat in a fragmented rainforest in Costa Rica. Primates 59, 301–311. https://doi.org/10.1007/s10329-018-0652-0
Brockerhoff, E.G., Barbaro, L., Castagneyrol, B., Forrester, D.I., Gardiner, B., González-Olabarria, J.R., et al., 2017. Forest biodiversity, ecosystem functioning and the provision of ecosystem services. Biodivers. Conserv. 26, 3005–3035. https://doi.org/10.1007/s10531-017-1453-2
Burnham, K.P., Anderson, D.R., 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, Second. ed. Springer US, New York.
Causey, O.R., Kumm, H.W, Laemmert Jr., H., 1950. Dispersion of Forest Mosquitoes in Brazil: Further Studies. Am. J. Trop. Med. Hyg. s1-30, 301–312. https://doi.org/10.4269/AJTMH.1950.S1-30.301
Charles, H., Dukes, J.S., 2007. Impacts of Invasive Species on Ecosystem Services, in: Biological Invasions. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 217–237. https://doi.org/10.1007/978-3-540-36920-2_13
Childs, M.L., Nova, N., Colvin, J., Mordecai, E.A., 2019. Mosquito and primate ecology predict human risk of yellow fever virus spillover in Brazil. Philos. Trans. R. Soc. B Biol. Sci. 374. https://doi.org/10.1098/rstb.2018.0335
Chippaux, J.P., Chippaux, A., 2018. Yellow fever in Africa and the Americas: a historical and epidemiological perspective. J. Venom. Anim. Toxins Incl. Trop. Dis. 24, 20. https://doi.org/10.1186/s40409-018-0162-y
Couto-Lima, Di., Madec, Y., Bersot, M.I., Campos, S.S., Motta, M.D.A., Santos, F.B. dos, et al., 2017. Potential risk of re-emergence of urban transmission of Yellow Fever virus in Brazil facilitated by competent Aedes populations. Sci. Rep. 7, 4848. https://doi.org/10.1038/s41598-017-05186-3
Crespin, S.J., Simonetti, J.A., 2016. Loss of ecosystem services and the decapitalization of nature in El Salvador. Ecosyst. Serv. 17, 5–13. https://doi.org/10.1016/j.ecoser.2015.10.020
Cunha, M. dos P., Duarte-Neto, A.N., Pour, S.Z., Ortiz-Baez, A.S., Černý, J., Pereira, B.B. de S., et al., 2019. Origin of the São Paulo Yellow Fever epidemic of 2017–2018 revealed through molecular epidemiological analysis of fatal cases. Sci. Rep. 9, 20418. https://doi.org/10.1038/s41598-019-56650-1
Danon, L., Ford, A.P., House, T., Jewell, C.P., Keeling, M.J., Roberts, G.O., et al., 2011. Networks and the Epidemiology of Infectious Disease. Interdiscip. Perspect. Infect. Dis. 2011, 1–28. https://doi.org/10.1155/2011/284909
Dantas-Torres F., Otranto, D., 2015. Vector-Borne Zoonoses., In: Sing A. (eds) Zoonoses-Infections Affecting Humans and Animals: Focus on Public Health Aspects. Springer Netherlands, Dordrecht. https://doi.org/10.1007/978-94-017-9457-2
Davidson, G., Chua, T.H., Cook, A., Speldewinde, P., Weinstein, P., 2019. The Role of Ecological Linkage Mechanisms in Plasmodium knowlesi Transmission and Spread. Ecohealth 16, 594–610. https://doi.org/10.1007/s10393-019-01395-6
de Almeida, M.A.B., dos Santos, E., Cardoso, J. da C., da Silva, L.G., Rabelo, R.M., Bicca-Marques, J.C., 2019. Predicting Yellow Fever Through Species Distribution Modeling of Virus, Vector, and Monkeys. Ecohealth 16, 95–108. https://doi.org/10.1007/s10393-018-1388-4
Fahrig, L., 2003. Effects of Habitat Fragmentation on Biodiversity. Annu. Rev. Ecol. Evol. Syst. 34, 487–515. https://doi.org/10.1146/annurev.ecolsys.34.011802.132419
Ferraguti, M., Martínez-de la Puente, J., Figuerola, J., 2021. Ecological Effects on the Dynamics of West Nile Virus and Avian Plasmodium: The Importance of Mosquito Communities and Landscape. Viruses 13, 1208. https://doi.org/10.3390/v13071208
Figueiredo, H.., 2017. Anthropic impacts on remaining urban forest Fragments: a new approach to environmental protection area monitoring. LAP LAMBERT Academic Publishing, Beau Bassin.
Fox, J., Weisberg, S., 2018. An R Companion to Applied Regression, Third Edit. SAGE Publications.
Gomes, A. de C., Torres, M.A.N., Paula, M.B. de, Fernandes, A., Marassá, A.M., Consales, C.A., et al., 2010. Ecologia de Haemagogus e Sabethes (Diptera: Culicidae) em áreas epizoóticas do vírus da febre amarela, Rio Grande do Sul, Brasil. Epidemiol. e Serviços Saúde 19, 101–113. https://doi.org/10.5123/S1679-49742010000200003
Gottdenker, N.L., Streicker, D.G., Faust, C.L., Carroll, C.R., 2014. Anthropogenic Land Use Change and Infectious Diseases: A Review of the Evidence. https://doi.org/10.1007/s10393-014-0941-z
Grace, D., Gilbert, J., Randolph, T., Kang, E., 2012. The multiple burdens of zoonotic disease and an ecohealth approach to their assessment 44, 67–73. https://doi.org/10.1007/s11250-012-0209-y
Hill, S.C., de Souza, R., Thézé, J., Claro, I., Aguiar, R.S., Abade, L., et al., 2020. Genomic Surveillance of Yellow Fever Virus Epizootic in São Paulo, Brazil, 2016–2018. PLOS Pathog. 16, e1008699. https://doi.org/10.1371/journal.ppat.1008699
IBGE, 2021. Instituto Brasileiro de Geografia e Estatistica: Cidades e Estados [WWW Document]. URL https://www.ibge.gov.br/cidades-e-estados/sp/ (accessed Jun 2 2022).
Ilacqua, R.C., Medeiros-Sousa, A.R., Ramos, D.G., Obara, M.T., Ceretti-Junior, W., Mucci, L.F., et al., 2021. Reemergence of Yellow Fever in Brazil: The Role of Distinct Landscape Fragmentation Thresholds. J. Environ. Public Health 2021, 1–7. https://doi.org/10.1155/2021/8230789
Instituto Florestal, 2021. Infraestrutura e Meio Ambiente: Reserva da Biosfera do Cinturão Verde da Cidade de São Paulo [WWW Document]. URL https://www.infraestruturameioambiente.sp.gov.br/institutoflorestal/o-instituto/rbcv/o-cinturao-verde/ (accessed Jun 21 2022).
Jackson, H.B., Fahrig, L., 2012. What size is a biologically relevant landscape? Landsc. Ecol. 27, 929–941. https://doi.org/10.1007/s10980-012-9757-9
Jones, B.A., Grace, D., Kock, R., Alonso, S., Rushton, J., Said, M.Y., 2013. Zoonosis emergence linked to agricultural intensi fi cation and environmental change 110, 8399–8404. https://doi.org/10.1073/pnas.1208059110
Jones, K.E., Patel, N.G., Levy, M.A., Storeygard, A., Balk, D., Gittleman, J.L., et al., 2008. Global trends in emerging infectious diseases. Nature 451, 990–993. https://doi.org/10.1038/nature06536
Le Flohic, G., Porphyre, V., Barbazan, P., Gonzalez, J.P., 2013. Review of Climate, Landscape, and Viral Genetics as Drivers of the Japanese Encephalitis Virus Ecology. PLoS Negl. Trop. Dis. 7, e2208. https://doi.org/10.1371/journal.pntd.0002208
Lenz, B.B., Jack, K.M., Spironello, W.R., 2014. Edge effects in the primate community of the biological dynamics of forest fragments project, Amazonas, Brazil. Am. J. Phys. Anthropol. 155, 436–446. https://doi.org/10.1002/ajpa.22590
Loh, E.H., Zambrana-Torrelio, C., Olival, K.J., Bogich, T.L., Johnson, C.K., Mazet, J.A.K., et al., 2015. Targeting Transmission Pathways for Emerging Zoonotic Disease Surveillance and Control. Vector-Borne Zoonotic Dis. 15, 432–437. https://doi.org/10.1089/vbz.2013.1563
Lopes, J., 1997. Ecologia de mosquitos (Diptera: Culicidae) em criadouros naturais e artificiais de área rural do Norte do Estado do Paraná, Brasil. V. Coleta de larvas em recipientes artificiais instalados em mata ciliar. Rev. Saude Publica 31, 370–377. https://doi.org/10.1590/S0034-89101997000400006
Loyal, J.D., Chen, Y., 2020. Statistical Network Analysis: A Review with Applications to the Coronavirus Disease 2019 Pandemic. Int. Stat. Rev. 88, 419–440. https://doi.org/10.1111/insr.12398
MapBiomas, 2020. Projeto Maobiomas: Coleção 5.0 da Série Anual de Mapas de Cobertura e Uso de Solo do Brasil, periodo em 1985 e 2019. [WWW Document]. MapBiomas. URL https://plataforma.mapbiomas.org/ (accessed Jun 10 2022).
Martínez-López, B., Perez, A.M., Sánchez-Vizcaíno, J.M., 2009. Social Network Analysis. Review of General Concepts and Use in Preventive Veterinary Medicine. Transbound. Emerg. Dis. 56, 109–120. https://doi.org/10.1111/j.1865-1682.2009.01073.x
Mbora, D.N.M., McPeek, M.A., 2009. Host density and human activities mediate increased parasite prevalence and richness in primates threatened by habitat loss and fragmentation. J. Anim. Ecol. 78, 210–218. https://doi.org/10.1111/j.1365-2656.2008.01481.x
MEA, 2005. Millennium Ecosystem Assessment: Ecosystems and Human Well-being: Synthesis. Island Press, Washington, DC.
Medeiros-Sousa, A.R., Fernandes, A., Ceretti-Junior, W., Wilke, A.B.B., Marrelli, M.T., 2017. Mosquitoes in urban green spaces: using an island biogeographic approach to identify drivers of species richness and composition. Sci. Rep. 7, 17826. https://doi.org/10.1038/s41598-017-18208-x
Mishra, J., Mishra, P., Arora, N.K., 2021. Linkages between environmental issues and zoonotic diseases: with reference to COVID-19 pandemic. Environ. Sustain. 4, 455–467. https://doi.org/10.1007/s42398-021-00165-x
Monath, T.P., Vasconcelos, P.F.C., 2015. Yellow fever. J. Clin. Virol. 64, 160–173. https://doi.org/10.1016/j.jcv.2014.08.030
Mori, A.S., 2017. Biodiversity and ecosystem services in forests: management and restoration founded on ecological theory. J. Appl. Ecol. 7–11. https://doi.org/10.1111/1365-2664.12854
MS, 2021. Ministério da Saúde: Boletim Epidemiológico - Situação epidemiológica da febre amarela – Monitoramento 2020/2021, Boletim Epidemiológico.
Mucci, F., Medeiros-sousa, A., Ceretti-Jr, W., Fernandes, A., Camargo, A., Evangelista, E., 2016. Haemagogus leucocelaenus and Other Mosquitoes Potentially Associated With Sylvatic Yellow Fever In Cantareira State Park In the São Paulo Metropolitan Area, Brazil. J. Am. Mosq. Control Assoc. 32, 329–332,.
Mucci, L.F., Júnior, R.P.C., de Paula, M.B., Scandar, S.A.S., Pacchioni, M.L., Fernandes, A., et al., 2015. Feeding habits of mosquitoes (Diptera: Culicidae) in an area of sylvatic transmission of yellow fever in the state of São Paulo, Brazil. J. Venom. Anim. Toxins Incl. Trop. Dis. 21, 6. https://doi.org/10.1186/s40409-015-0005-z
Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation priorities. Nature 403, 853–858. https://doi.org/10.1038/35002501
Ogrzewalska, M., Uezu, A., Jenkins, C.N., Labruna, M.B., 2011. Effect of Forest Fragmentation on Tick Infestations of Birds and Tick Infection Rates by Rickettsia in the Atlantic Forest of Brazil. Ecohealth 8, 320–331. https://doi.org/10.1007/s10393-011-0726-6
Ostfeld, R.S., Keesing, F., 2000. Biodiversity and Disease Risk: the Case of Lyme Disease. Conservation Biology 14, 722–728
Plowright, R.K., Parrish, C.R., Mccallum, H., Hudson, P.J., Ko, A.I., Graham, A.L., Lloyd-smith, J.O., 2017. Pathways to zoonotic spillover. Nat Rev Microbiol. 15, 502–510. https://doi.org/10.1038/nrmicro.2017.45
Plowright, R.K., Reaser, J.K., Locke, H., Woodley, S.J., Patz, J.A., Becker, D.J., et al., 2021. Land use-induced spillover: a call to action to safeguard environmental, animal, and human health. Lancet Planet. Heal. 5, e237–e245. https://doi.org/10.1016/S2542-5196(21)00031-0
Possas, C., Lourenço-de-Oliveira, R., Tauil, P.L., Pinheiro, F. de P., Pissinatti, A., Cunha, et al., 2018. Yellow fever outbreak in Brazil: the puzzle of rapid viral spread and challenges for immunisation. Mem. Inst. Oswaldo Cruz 113, e180278. https://doi.org/10.1590/0074-02760180278
Prist, P.R., D´Andrea, P.S., Metzger, J.P., 2017. Landscape, Climate and Hantavirus Cardiopulmonary Syndrome Outbreaks. Ecohealth 14, 614–629. https://doi.org/10.1007/s10393-017-1255-8
Prist, P.R., Prado, A., Tambosi, L.R., Umetsu, F., de Arruda Bueno, A., Pardini, R., et al., 2021a. Moving to healthier landscapes: Forest restoration decreases the abundance of Hantavirus reservoir rodents in tropical forests. Sci. Total Environ. 752, 141967. https://doi.org/10.1016/j.scitotenv.2020.141967
Prist, P., Tambosi, L., Mucci, L., Pinter, A., de Souza, R., Muylaert, R., et al., 2021b. Roads and forest edges facilitate yellow fever virus dispersion. J. Appl. Ecol. 1–14. https://doi.org/10.1111/1365-2664.14031
Prist, P.R., Uriarte, M., Tambosi, L.R., Prado, A., Pardini, R., D’Andrea, P.S., et al., 2016. Landscape, environmental and social predictors of Hantavirus risk in São Paulo, Brazil. PLoS One 11, 1–18. https://doi.org/10.1371/journal.pone.0163459
Reed, S.C., Reibold, R., Cavaleri, M.A., Alonso-Rodríguez, A.M., Berberich, M.E., Wood, T.E., 2020. Soil biogeochemical responses of a tropical forest to warming and hurricane disturbance, in: Advances in Ecological Research. Elsevier Ltd., pp. 225–252. https://doi.org/10.1016/bs.aecr.2020.01.007
Rezende, C.L., Scarano, F.R., Assad, E.D., Joly, C.A., Metzger, J.P., Strassburg, B.B.N., et al., 2018. From hotspot to hopespot: An opportunity for the Brazilian Atlantic Forest. Perspect. Ecol. Conserv. 16, 208–214. https://doi.org/10.1016/j.pecon.2018.10.002
Runghen, R., Poulin, R., Monlleó-Borrull, C., Llopis-Belenguer, C., 2021. Network Analysis: Ten Years Shining Light on Host–Parasite Interactions. Trends Parasitol. 37, 445–455. https://doi.org/10.1016/j.pt.2021.01.005
Schmidt, K.A., Ostfeld, R., 2001. Biodiversity and the Dilution Effect in Disease Ecology. Ecol. Soc. Am. 609–619. https://doi.org/https://doi.org/10.1890/00129658(2001)082[0609:BATDEI]2.0.CO;2
Scinachi, C.A., Takeda, G.A.C.G., Mucci, L.F., Pinter, A., 2017. Association of the occurrence of Brazilian spotted fever and Atlantic rain forest fragmentation in the São Paulo metropolitan region, Brazil. Acta Trop. 166, 225–233. https://doi.org/10.1016/j.actatropica.2016.11.025
Silva, M.A.N. da, Lozovei, A.L., 1999. Ocorrência de Haemagogus (Conopostegus) leucocelaenus (Dyar & Shannon) e Toxorhynchites (Lynchiella) theobaldi (Dyar & Knab) em ocos de árvore em capão de mata, Curitiba, Paraná, Brasil. Rev. Bras. Zool. 16, 257–267. https://doi.org/10.1590/S0101-81751999000500018
SOSMA, 2020. Qual é a área de cobertura da Mata Atlântica? [WWW Document]. Fundação SOS Mata Atl. URL https://www.sosma.org.br/artigos/qual-e-area-de-cobertura-da-mata-atlantica/ (accessed Jun 12 2022).
Trevelin, L.C., Port-Carvalho, M., Silveira, M., Morell, E., 2007. Abundance, habitat use and diet of Callicebus nigrifrons Spix (Primates, Pitheciidae) in Cantareira State Park, São Paulo, Brazil. Rev. Bras. Zool. 24, 1071–1077. https://doi.org/10.1590/S0101-81752007000400026
Venables WN, R.B., 2002. Modern Applied Statistics with S., Fourth Edit. Springer, New York.
WHO, 2021. Zoonoses Key facts [WWW Document]. URL https://www.who.int/news-room/fact-sheets/detail/zoonoses (accessed Jun 20 2022).
Wickham, H., 2016. ggplot2: Elegant Graphics for Data Analysis, Second Edit. Springer, New York. https://doi.org/10.1007/978-3-319-24277-4
Wilk-da-Silva, R., Medeiros-Sousa, A.R., Laporta, G.Z., Mucci, L.F., Prist, P.R., Marrelli, M.T., 2022. The influence of landscape structure on the dispersal pattern of yellow fever virus in the state of São Paulo. Acta Trop. 228, 106333. https://doi.org/10.1016/j.actatropica.2022.106333
Wilk-da-Silva, R., Mucci, L.F., Ceretti-Junior, W., Duarte, A.M.R. de C., Marrelli, M.T., Medeiros-Sousa, A.R., 2020. Influence of landscape composition and configuration on the richness and abundance of potential sylvatic yellow fever vectors in a remnant of Atlantic Forest in the city of São Paulo, Brazil. Acta Trop. 204, 105385. https://doi.org/10.1016/j.actatropica.2020.105385
Zittra, C., Vitecek, S., Obwaller, A.G., Rossiter, H., Eigner, B., Zechmeister, T., et al., 2017. Landscape structure affects distribution of potential disease vectors (Diptera: Culicidae). Parasit. Vectors 10, 205. https://doi.org/10.1186/s13071-017-2140-6
Zohdy, S., Schwartz, T.S., Oaks, J.R., 2019. The Coevolution Effect as a Driver of Spillover. Trends Parasitol. 35, 399–408. https://doi.org/10.1016/j.pt.2019.03.010

No competing interests reported.

SupplementaryFig1.tif
Supplementary Fig. 1. Map of the state of São Paulo. Municipalities with records of epizootics are shown in pink, and neighboring municipalities without occurrences in blue. Purple lines correspond to links between municipalities with records of occurrences separated by no more than 30 km between the centroids and with a period of no more than 30 days between cases of epizootics in NHPs. Circles correspond to the overall total of links for each municipality; the municipalities with the greatest number of links are marked with a fluorescent green star.
SupplementaryFig2.tiff
Supplementary Fig. 2. Landscape metrics associated with forest fragments in municipalities in the state of São Paulo with records of epizootics. A) Euclidean Nearest Neighbor Distance – ENN (distance to the nearest neighboring patch). B) Edge Density – ED (m/ha); Largest Patch Index - LPI (proportion of the landscape comprised by the largest patch); Patch Density – PD (number per 100 ha); Percentage of Landscape – PLAND (%); Percentage of Like Adjacencies - PLADJ (number of like adjacencies involving the Forest core areas).
SupplementaryTable1.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

The role of deforestation and fragmentation in yellow fever virus dispersal

Status:

Version 1

Abstract

Figures

1. Introduction

2. Materials And Methods

2.1. Study area

2.2. Data on epizootics

2.3. Landscape data

2.4. Network analysis

2.5. Statistical analysis

3. Results

4. Discussion

5. Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1