Guidelines for Arbovirus Surveillance Programs in the United States
Click here for CONTENTS Page
For contact details regarding publication see Title Page

Introduction
Meteorologic Data Monitoring
Vector Surveillance
Aedes melanimon
Culex tarsalis
Vertebrate Host Surveillance

Introduction

WEE virus occurs from about the Mississippi River west to the Pacific coast, (Fig. 6-1) including the prairie provinces of Canada ⁶ and the western states of Mexico. It occasionally produces epizootics and epidemics, but regularly causes equine and human cases ²⁴¹. Although WEE virus was previously thought to occur nationwide, it was subsequently discovered that the agent in the east was a separate virus, which was renamed Highlands J (HJ).¹⁵⁰ HJ virus is rarely pathogenic for horses, and is not known to be pathogenic for humans.

Figure 6-1. Geographic distribution of confirmed and presumptive human cases of western equine encephalomyelitis in the United States 241.

Epidemics of WEE recur at irregular intervals or from 10 to 11 years (Fig. 6-2) For human cases reported for the period 1955 through 1992, autocorrelation analysis shows a recurrence of major activity approximately every 10 years. Reiter ²⁴⁷ has discussed several climatic factors that could lead to cyclic recrudescence of viruses such as WEE (Also, see below).

Meteorologic Data Monitoring

The delayed accumulation of 50 degree days above 70° F, indicating a long cool spring, has been associated with increased WEE virus transmission.¹³⁰ The date of temperature inversion in soil was shown to correlate with the occurrence of Cx. tarsalis-borne WEE in humans and horses. In years of heavy snowmelt runoff or increased spring precipitation, flooding may create more larval habitats for vector species such as Cx. tarsalis, Cs. inornata, and Aedes spp. Prolonged cool and wet weather in spring also may increase mosquito survival. Long-lived females are more likely to become infected and transmit virus. Snowpack measurements by themselves have been variably associated with epidemic WEE transmission.

Figure 6-2. Reported cases of confirmed and presumptive human cases of western equine encephalomyelitis in the United States, 1964-1992.

Elevated temperatures in midsummer have been associated with diminished activity of adult Cx. tarsalis mosquitoes; in California, this leads to reduced abundance in light trap collections in the Coachella and Imperial Valleys during August and September.²¹⁹ Infected adult females modulate their infections through prolonged hot periods, reducing transmission efficiency.¹¹⁶ The relative importance of modulation and adult mortality as reducers of transmission have not been studied under field conditions. Retrospective analysis of cases in three epidemic years showed that the hottest weeks of the summer were followed by a decline in epizootic transmission. With the return of cooler temperatures, transmission resumed at a high level.^l See Chapter 5 for an additional discussion weather and climate effects on Cx. tarsalis-transmitted arboviruses.

In a study comparing 2 epidemic and 2 non-epidemic years, the timing and location of WEE outbreaks in horses and humans, seroconversions in sentinel chickens, and first isolation of WEE virus mark-from Cx. tarsalis could be correlated with wind trajectories from states further south.²⁵⁶ It remains to be demonstrated whether there is a causal relationship between weather fronts and the appearance of WEE virus and cases.

General information on the biology, behavior and control of WEE vectors is available in separate publications. ^{49,189,233,241}

Aedes melanimon: (No common name ). ²⁸¹

In the Sacramento Valley of California, Ae. melanimon is involved in a secondary transmission cycle of WEE involving jackrabbits.^145,235 This species has been reported from California, Oregon, Washington, Nevada, Utah, Idaho, Montana, Wyoming, Colorado and New Mexico, and from Alberta, Canada.

A combination of spring flooding, warming temperatures and increasing daylength stimulate eclosion of Ae. melanimon eggs. Larvae are commonly associated with irrigated pasture and waterfowl areas. In brackish water habitats, Ae. melanimon is replaced by Ae. dorsalis.²⁴³ Ae. melanimon is multivoltine and, depending on water level fluctuations in larval habitats, can produce up to 12 or more broods per season.²⁴³

Peak flight activity occurs during the twilight hours in the spring and summer. However, nocturnal flight activity may increase during the fall. Aedes melanimon females are strong fliers. They may disperse 8 to 10 miles or more from breeding sites, particularly when aided by prevailing winds. Morning peaks in flight activity are probably associated with searches for resting sites rather than host-seeking and feeding.²⁴³

Aedes melanimon readily bites humans, and the species is a major pest in some areas. Leporids (hares and rabbits) serve as principal hosts. Other hosts include cattle, horses, sheep, deer and dogs. This species seldom feeds upon birds.²⁴³ The females will bite during the day if disturbed. However, biting activity occurs primarily in the first 2 hours after sunset. There is evidence that parous females feed slightly later in the evening than nulliparous females.²⁴³

Daily survival has been estimated for this species in the Sacramento Valley of California.¹⁴⁵ Survivorship was estimated at 0.84 to 0.90 in mark-from release-recapture studies, 0.82 to 0.89 in parity state studies. Another study found that about 4% and 1% of 319 specimens had completed 2 and 3 or more gonotrophic cycles, respectively.¹⁸⁴ Adults can be collected in large numbers in CO2 -baited CDC light traps. However, older females may be more frequently collected in New Jersey light traps.¹⁸⁴ This species is not readily collected from resting boxes.¹⁸⁴

Culex tarsalis: (No common name ).²⁸¹

Culex tarsalis is the primary enzootic, epizootic and epidemic vector of WEE virus in the United States.^241,243 For practical purposes WEE virus surveillance in mosquitoes can be limited to the collecting and testing of Cx. tarsalis. Occasional WEE virus isolates may be obtained from other mosquito species collected concurrently, or sometimes earlier in the season. The significance of such findings and their relationship to WEE virus activity are unknown.

Culex tarsalis is found from western Canada, through the United States, south to the state of Chiapas, Mexico. In Canada there are records from British Columbia, Alberta, Saskatchewan, Manitoba, and the Northwest Territories.⁴⁶ In the United States Cx. tarsalis is generally common west of the Mississippi River. It is usually uncommon or rare in the eastern part of the country. However, it has been collected as far east as New Jersey and Rhode Island.^65,142 The distribution of Cx. tarsalis shows focal clustering in the Great Plains, prairie, and other grassland areas. The vertical distribution of Cx. tarsalis extends from below sea level to almost 10,000 feet in California.³¹

Larval habitats of Cx. tarsalis are closely associated with irrigated farm and ranch lands.¹⁸⁶ In Kern County, California, temporary to semi-permanent earth-lined sites were the preferred larval habitat in 48% of 860 collections of this species. Only 13% of the collections came from artificially-lined containers.²⁴³ Open, unshaded sites were preferred over shaded sites. Irrigation water, especially waste tailwater, was the most common source of larval habitats.²⁴³

During daylight hours the adults rest in secluded spots. A variety of natural habitats serve as resting sites. These include animal burrows, grass and shrubs, artificial shelters such as the underside of bridges. Privies, culverts, cellars, chicken houses, and other farm buildings also may serve as resting sites. Light, temperature, and relative humidity are important variables that determine the suitability of such sites.

The seasonal abundance and duration of annual activity of Cx. tarsalis are influenced by latitude and temperature. Throughout much of its range the maximum adult population is reached during August or September. However, population peaks usually occur during May-June in Imperial and Coachella Valleys of southern California. In the Central Valley of California peaks have occurred in May-June, but more typically occur in July-September. Peaks have been recorded as early as July in Washington and in Alberta, Canada. Most collection records for Cx. tarsalis east of the Mississippi River are in late autumn. This species occurs in the Tennessee Valley from late August to late November, with a population peak in September. In west Texas Cx. tarsalis is abundant from June through September. Farther south in the Lower Rio Grande Valley, Cx. tarsalis is most abundant during November and occurs throughout the winter in appreciable numbers. Populations then begin to decline and few specimens are collected during April and May, and none from June through September. A similar situation occurs in the extreme southern valleys of California.^49,219,243

Adults are active chiefly from dusk to dawn, with peak activity occurring within 2 hours after sunset. In a study using truck traps in Kern County, California, males were found to leave diurnal resting sites first. Males were followed by empty, blooded and gravid females, respectively.²⁴³ Adults began returning shortly before sunrise, and entry into resting sites was in the reverse order of leaving. It is believed that most Cx. tarsalis females remain within 50 feet of the ground in flight,¹⁰ although this species has been collected as high as 610 m (2,000 ft) over central Texas.¹⁰⁵ Dispersal occurs in all directions at low wind velocities, but mosquitoes orient into the wind as velocities increase. Winds more than 6 mph inhibit flight. Culex tarsalis females can travel 8 to 10 miles -in 2 evenings. They may spread as far as 25 miles from breeding sites.¹⁰

Culex tarsalis feeds readily on humans out-of- doors during the summer months. Peak human-biting activity usually begins about 30 minutes after sunset and lasts for about l hour. Human avoidance of exposure to mosquito bites during the first couple of hours after sunset can be a practical preventive measure during the WEE transmission season. However, bites received in the early morning may have a higher probability of being infective because of increased parity among females feeding then.²⁴³

Precipitin test studies have shown that Cx. tarsalis is a general feeder with a preference for avian hosts in most areas during certain seasons of the year.²⁸⁴ Culex tarsalis may feed almost exclusively on birds in the spring, but during the summer increasing numbers of females also feed on mammalian hosts. This shift in the feeding pattern often coincides with the appearance of WEE virus infection in humans and other vertebrates. It may be an important factor making Cx. tarsalis such an efficient enzootic, epizootic and epidemic vector. The reasons for the observed seasonal shift in the feeding pattern have not been fully elucidated. However, host availability, host defensive reactions mosquito density, and other seasonal variables may all play a role.²⁴³

Inseminated females may seek a blood meal, or in some cases may develop the first egg batch autogenously (i.e., without benefit of a blood meal).¹⁷ The proportion of autogeny varies seasonally.^198,270 Anautogenous females will take a blood meal as early as the third day after emergence under laboratory conditions, and oviposit 4 days later. In the Central Valley of California, Cx. tarsalis can complete development during the summer in irrigated pastures within 9 to 10 days following irrigation.

Daily survival rates for Cx. tarsalis in Kern County, California have been estimated by constructing both vertical and horizontal life tables. Estimates were made at two sites from May through September over several years.²⁴³ Seasonal mean survival rates varied from 0.63 to 0.86 per day. Estimates tended to be lower in July, possibly due to dilution by newly-emerged adults. In the Sacramento Valley of California, an emergence-independent vertical method estimated daily survival at 0.86 and 0.84 for empty and blood-fed females, respectively.¹⁷³

Culex tarsalis females can be collected by a variety of methods. New Jersey light traps or CO2 -in baited CDC light traps are effective, as are lard-can bait traps using either chickens or dry ice as bait. Walk-in or cubic-foot resting boxes can be used to collect resting females, as can aspirator collections from culverts, bridges, chicken houses, etc. In California, New Jersey light trap indices have been used to establish thresholds for virus transmission in urban and rural environments.²²⁴ In a single California study, the Reiter gravid trap²⁴⁴ was not effective in collecting Cx. tarsalis. ²⁴²

The ecology of WEE consists of a wild bird- Cx. tarsalis cycle throughout the irrigated portion of western North America and along waterways in the northern plains states. Although WEE virus has been isolated from other vertebrates (rodents, jackrabbits and reptiles) and from other vectors (Culiseta inornata and Aedes spp.), only a few species of passerine birds free-and the principal vector, Cx. tarsalis, are responsible for summer amplification.²³³

The density and availability of susceptible bird species (particularly nestlings), vector density and their temporal and spatial interaction are important factors in the summer amplification of WEE. The early amplification of WEE virus transmission within the bird-mosquito cycle will increase the proportion of infected adult mosquitoes in the population. Since Cx. tarsalis normally shifts its host-seeking from birds to mammals in midsummer ^232,232 this higher infection ratio increases the probability of transmission of WEE to mammals when the mosquito shifts its host-feeding behavior. This increases the risk to equine and human populations.

Various measures of early viral activity have been employed to predict the occurrence of WEE cases and outbreaks. These include virus in wild avian hosts, sentinel chickens, equines or mosquito vectors, and the abundance of mosquito vectors. Monitoring WEE viral infections in birds locally involved in early amplification provides valuable information about the amount and extent of early viral transmission. This can help determine impending risk. Studies in west Texas in 1965-1969¹³³ demonstrated that WEE viral activity in nestling house sparrows and in Cx. tarsalis started by mid to late June. Activity continued in house sparrows for 8-10 weeks and in Cx. tarsalis for 12-13 weeks. A similar temporal pattern of virus activity was observed in North Dakota in 1975.¹⁷⁹ Serologic surveys in Kern County, California, found higher HI antibody prevalences against WEE virus in winter months, but WEE virus isolations were obtained from nestling birds from mid June to mid August.²³³

Surveillance programs for WEE virus vary because of differences in 1) professional orientation of the investigators, 2) ecology of vertebrate hosts and mosquito vectors, and 3) climate, physiography and agricultural practices. In Kern County, California, the birds with the highest antibody prevalence during epidemics were the house finch, house sparrow, blackbirds, orioles and mourning dove. Nestling house finches and pigeons were also valuable indicators when available.²³⁴ Sentinel chickens were used to detect movement of WEE virus from enzootic foci to peridomestic settings before equine or human cases. A comparative study in California concluded that pigeons were less suitable than chickens as sentinels.²³⁸

In west Texas, infection rates in house sparrows were the best predictors of human disease.^120,133 This was true for antibody rates in free-ranging birds and for viremia in nestlings. Virus isolation rates of 5-6% in nestlings and antibody rates of 45-56% in free-ranging birds were common.¹³³ House sparrows were singularly useful in that area of Texas. They constituted more than two-thirds of the local avian population, were closely associated with humans and the vector mosquito, and were quite accessible for sampling.

In the northern plains states, other avian species had higher antibody prevalences and were equal in abundance and accessibility. In North Dakota house sparrows, the antibody prevalence was 13% and no virus isolations were obtained from nestlings. In contrast, there was a 46% antibody rate in the American robin. There were nine isolations of WEE virus, including seven from nestlings of four species other than house sparrows.¹⁷⁹ In Colorado during 1987, the antibody prevalences were 8% in house sparrows, 29% in American robins, 21% in black-hosts, capped chickadees, 15% in pigeons, 9% in red-winged blackbirds, and 7% in waterfowl.^m

Seroconversions in sentinel chickens and equine cases have been used to monitor WEE virus activity for decades.²³³ The advantages and disadvantages of using them are presented elsewhere in this publication (See Ch.1).

Gaps in current knowledge of western equine encephalitis

• What are the most reliable predictors for human risk of WEE infection?
• What predictors for WEE viral activity can be used in the Rocky Mountain and Great Plains regions?
• Are there any large-scale regional predictors for WEE viral activity?
• What is the most effective way to control vectors of WEE in an emergency (e.g., widespread flooding)?
• How can we improve surveillance for cases in humans and equines?
• Why are there few human or equine cases of WEE along the lower Colorado River in the presence of high seroconversion rates in chickens and numerous isolates from Cx. tarsalis?
• What is the overwintering mechanism of WEE virus?
• What is the role of wind in the dispersal of WEE vectors over regional (i.e., > 100 km) distances?
• Are there other host-vector cycles for WEE virus (e.g., Ae. melanimon - jackrabbit cycle) outside California?
• Can ovarian dissection or other age-determination procedures give a more accurate estimate of the likelihood of WEE virus transmission, as with EEE in New Jersey? How does autogeny impact upon 64 parity estimates?
• Are there enzootic and epizootic/epidemic strains of WEE virus that have differing ecologies? ¹

^m McLean, R.G., Unpublished data.