• View in gallery

    Deviation from the long-term average of the number of days with a daily maximum (max) temperature approximately 27°C versus the day of year of crossover from Culex restuans to Cx. pipiens. The number of days is accumulated from January 1.

  • View in gallery

    Square of the correlation coefficient between crossover date and accumulated degree days as a function of degree day base for data from the period 1988–2003 (except for 1989, 1993, and 2000). Degree days are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

  • View in gallery

    Square of the correlation coefficient between cross-over date and accumulated number of maximum temperature (Tmax) exceedances as a function of exceedance threshold for data from the period 1988–2003 (except for 1989, 1993, and 2000). Threshold exceedances are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

  • View in gallery

    Square of the correlation coefficient between cross- over date and accumulated number of minimum temperature (Tmin) exceedances as a function of exceedance threshold for data from the period 1988–2003 (except for 1989, 1993, and 2000). Threshold exceedances are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

  • View in gallery

    Time series of the day of year of crossover from Culex restuans to Cx. pipiens for 1988–2003 for observed (Obs) data (thick solid line), estimates from a degree day model (DD Mod) using base temperature of 17°C (dashed line), and estimates from a maximum temperature (Tmax) exceedances model (Tmax Mod) using a thresh-old of 27°C (thin solid line).

  • View in gallery

    Probability distribution of estimated crossover date as a function of the date on which the estimation was made. The probability distribution was computed from 104 scenarios, each scenario representing one year from the period 1900–2003. Five values of the probability distribution are displayed: earliest, 90% probability of exceedance, median, 10% probability of exceedance, and latest.

  • 1

    Centers for Disease Control, 2004. West Nile Virus. Cited January 20, 2005. Available from http://www.cdc.gov/ncidod/dvbid/westnile/qa/cases.htm

  • 2

    Illinois Department of Public Health, 2005. West Nile Virus. 2003. Cited January 20, 2005. Available from http://www.idph.state.il.us/envhealth/wnv.htm

  • 3

    Hayes CG, 2001. West Nile virus: Uganda, 1937, to New York City, 1999. Ann N Y Acad Sci 951 :25–37.

  • 4

    Day JF, 2001. Predicting St. Louis encephalitis virus epidemics: lessons from recent, and not so recent, outbreaks. Annu Rev Entomol 46 :111–138.

    • Search Google Scholar
    • Export Citation
  • 5

    Tsai TF, Mitchell CJ, 1989. St. Louis encephalitis. Monath TP, ed. The Arboviruses: Epidemiology and Ecology. Boca Raton, FL: CRC Press, 113–143.

  • 6

    Bernard K, Kramer L, 2001. West Nile virus activity in the United States, 2001. Viral Immunol 14 :319–338.

  • 7

    Turrell MJ, O’Guinn M, Oliver J, 2000. Potential for New York mosquitoes to transmit West Nile virus. Am J Trop Med Hyg 62 :413–414.

  • 8

    Turrell MJ, O’Guinn M, Dohm DJ, Jones JW, 2001. Vector competence of North American mosquitoes (Diptera: Culicidae) for West Nile virus. J Med Entomol 38 :130–134.

    • Search Google Scholar
    • Export Citation
  • 9

    Bailey CL, Faran ME, Gargan TP II, Hayes DE, 1982. Winter survival of blood fed and non-blood-fed Culex pipiens. Am J Trop Med Hyg 31 :1054–1061.

    • Search Google Scholar
    • Export Citation
  • 10

    Mitchell CJ, 1988. Occurrence, biology, and physiology of diapause in overwintering mosquitoes. Monath TP, ed. The Arboviruses: Epidemiology and Ecology. Volume I. Boca Raton, FL: CRC Press, 191–217.

  • 11

    Redio J, Chen M-H, Meola R, 1999. Juvenile hormone biosynthesis in diapausing and nondiapausing Culex pipiens (Diptera: Culicidae). J Med Entomol 36 :355–360.

    • Search Google Scholar
    • Export Citation
  • 12

    Vinogradova EB, 2000. Characteristics of natural populations of Culex p. pipiens in Russia and the neighbouring countries. Golovatch SI, ed. Culex pipiens pipiens Mosquitoes: Taxonomy, Distribution, Ecology, Physiology, Genetics, Applied Importance and Control. Sofia, Bulgaria: Pensoft Publishers, 116–149.

  • 13

    Lampman RL, Novak RJ, 1996. Oviposition preference of Culex pipiens and Culex restuans for infusion-baited traps. J Am Mosq Control Assoc 12 :23–32.

    • Search Google Scholar
    • Export Citation
  • 14

    Coraso BG, Munstermann LE, 1984. Identification by electrophoresis of Culex adults (Diptera:Culcidae) in light-trap samples. J Med Entomol 21 :648–655.

    • Search Google Scholar
    • Export Citation
  • 15

    Lee JH, Rowley WA, 2000. The abundance and seasonal distribution of Culex mosquitoes in Iowa during 1995–97. J Am Mosq Control Assoc 16 :275–278.

    • Search Google Scholar
    • Export Citation
  • 16

    Lampman R, Hanson, Novak R, 1997. Seasonal abundance and distribution of mosquitoes at a rural waste tire site in Illinois. J Am Mosq Control Assoc 13 :193–200.

    • Search Google Scholar
    • Export Citation
  • 17

    Luby JP, Sulkin SE, Sanford JP, 1969. The epidemiology of St. Louis encephalitis: a review. Annu Rev Med 20 :329–350.

  • 18

    Mitchell CJ, Francy DP, Monath TP, 1980. Arthropod vectors. Monath TP, ed. St. Louis Encephalitis. Washington, DC: American Public Health Association, 313–379.

  • 19

    Helson BV, Surgeoner GA, Wright RE, 1980. The seasonal distribution and species composition of mosquitoes (Diptera: Culicidae) collected during a St. Louis Encephalitis Surveillance Program from 1976–1978 in southwestern Ontario, Canada. Can Entomol 112 :865–874.

    • Search Google Scholar
    • Export Citation
  • 20

    Monath TP, 1980. Epidemiology. Monath TP, ed. St. Louis encephalitis. Washington, DC: American Public Health Association, 239–312.

  • 21

    Campbell GL, Marfin AA, Lanciotti RS, Gubler DJ, 2002. West Nile virus. Lancet Infect Dis 2 :519–529.

  • 22

    Gubler DJ, Campbell GL, Nasci R, Komar N, Petersen L, Roehrig JT, 2000. West Nile virus in the United States: guidelines for detection, prevention, and control. Viral Immunol 13 :469–475.

    • Search Google Scholar
    • Export Citation
  • 23

    Madder DJ, Surgeoner GA, Helson BV, 1983. Number of generations, egg production, and developmental time of Culex pipiens and Culex restuans (Diptera: Culicidae) in southern Ontario (Canada). J Med Entomol 20 :275–287.

    • Search Google Scholar
    • Export Citation
  • 24

    Baker DG, Sharratt BS, Chiang HC, Zandlo JA, Ruschy DL, 1984. Base temperature selection for the prediction of European corn borer instars by the growing degree-day method. Agricult Forestr Meteorol 32 :55–60.

    • Search Google Scholar
    • Export Citation
  • 25

    Reisen W, Lothrop H, Chiles R, Madon M, Cossen C, Woods L, Husted S, Kramer V, Edman J, 2004. West Nile virus in California. Emerg Infect Dis 10 :1369–1378.

    • Search Google Scholar
    • Export Citation
  • 26

    Turell MJ, Dohm DJ, Sardelis MR, O Guinn ML, Andreadis TG, Blow JA, 2005. An update on the potential of North American mosquitoes (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol 42 :57–62.

    • Search Google Scholar
    • Export Citation
  • 27

    Kilpatrick AM, Kramer LD, Campbell SR, Alleyne EO, Dobson AP, Daszak P, 2005. West Nile virus risk assessment and the bridge vector paradigm. Emerg Infect Dis 11 :425–429.

    • Search Google Scholar
    • Export Citation
  • 28

    Monath TP, Tsai TF, 1987. St. Louis encephalitis: lessons from the last decade. Am J Trop Med Hyg 37 (Suppl 3):40S–59S.

  • 29

    Hubalek Z, 2000. European experience with the West Nile virus ecology and epidemiology: could it be relevant for the new world? Viral Immunol 13 :415–426.

    • Search Google Scholar
    • Export Citation
  • 30

    Ward MP, Levy M, Thacker HL, Ash M, Norman SKL, Moore GE, Webb PW, 2004. Investigation of an outbreak of encephalomyelitis caused by West Nile virus in 136 horses. J Am Vet Med Assoc 225 :84–89.

    • Search Google Scholar
    • Export Citation
  • 31

    Dohm DJ, O’Guinn ML, Turell MJ, 2002. Effect of environmental temperature on the ability of Culex pipiens (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol 39 :221–225.

    • Search Google Scholar
    • Export Citation
  • 32

    Reeves WC, Hammon WM, Longshore WA Jr, McClure HE, Geid AF, 1962. Epidemiology of the Arthropod-Borne Viral Encephalitides in Kern County, California, 1943–1952, Volume 4. Public Health, Berkeley, CA: University of California.

  • 33

    Hess AD, Cherubin CE, LaMotte LC, 1963. Relation of temperature to activity of western and St. Louis encephalitis viruses. Am J Trop Med Hyg 12 :657.

    • Search Google Scholar
    • Export Citation
  • 34

    Apperson CS, Harrison BA, Unnasch TR, Hassan HK, Irby WS, Savage HM, Aspen SE, Watson DW, Rueda LM, Engber BR, Nasci RS, 2002. Host-feeding habits of Culex and other mosquitoes (Diptera: Culicidae) in the Borough of Queens in New York City, with characters and techniques for identification of Culex mosquitoes. J Med Entomol 39 :777–785.

    • Search Google Scholar
    • Export Citation
  • 35

    Madder DJ, Macdonald RS, Surgeoner GA, Helson BV, 1980. The use of oviposition activity to monitor populations of Culex pipiens and Culex restuans (Diptera: Culicidae). Can Entomol 112 :1013–1018.

    • Search Google Scholar
    • Export Citation
  • 36

    Williams DD, Tavares CA, Kushiner DJ, Coleman JR, 1993. Colonization patterns and life-history dynamics of Culex mosquitoes in artificial ponds of different character. Can J Zool 71 :568–578.

    • Search Google Scholar
    • Export Citation
  • 37

    Duchon CE, 1986. Corn yield prediction using climatology. J Appl Meteor 25 :581–590.

  • 38

    Novak RJ, Lampman RL, 2001. Public health pesticides. Handbook of Pesticide Toxicology. Volume 1. Pesticide Risk Characterization. San Diego, CA: Academic Press, 181–201.

 

 

 

 

MODELING THE IMPACT OF VARIABLE CLIMATIC FACTORS ON THE CROSSOVER OF CULEX RESTAUNS AND CULEX PIPIENS (DIPTERA: CULICIDAE), VECTORS OF WEST NILE VIRUS IN ILLINOIS

View More View Less
  • 1 Illinois State Water Survey, Champaign, Illinois; Medical Entomology Program, Illinois Natural History Survey, Champaign, Illinois

The aim of this study was to model the impact of temperature on the timing of the seasonal shift in relative proportion of Culex restuans Theobald and Culex pipiens L. in Illinois. The temporal pattern of West Nile virus (WNV) and St. Louis encephalitis virus transmission in the midwest exhibits a late summer to early fall peak in activity, which parallels the temporal increase in the abundance of Cx. pipiens. The daily number of egg rafts oviposited by each species has been monitored at multiple surveillance sites in Urbana-Champaign in central Illinois for more than 13 years. The time when the two Culex species are in equal abundance (crossover) varies considerably from year to year. Our investigation of several thermal measures indicated that this variation was related in large part to climatic conditions with warmer (cooler) temperatures correlated to earlier (later) crossover dates. Models based on degree days and the number of days in which the daily maximum temperature exceeded an upper temperature threshold explained more than 60% of the variance in crossover dates. In contrast, models based on the number of days in which the daily minimum temperature exceeded a lower temperature threshold explained no more than 52% of the variance. An evaluation of these models demonstrated that they provide relatively simple and accurate estimates of crossover date from daily temperature data, a necessary component for developing an overall climatic index for the risk of WNV transmission in Illinois.

INTRODUCTION

West Nile Virus (WNV) is a mosquitoborne flavivirus typically transmitted between birds and mosquitoes, and endemic to Africa, Europe, the Middle East, west and central Asia, and Oceania. The virus was first identified in the United States in the New York metropolitan area in the fall of 1999, and since then has emerged as a threat to public, equine, and wildlife health in North America. In 2002, WNV emerged as a full-scale epidemic in the United States, being reported in 2,531 counties in 44 states, compared with 359 counties in 27 states in 2001. In addition, humans were much more heavily impacted in 2002 than in previous years with a total of 4,156 human cases and 284 related deaths reported. In contrast, 149 cases and 18 deaths were detected during the entire period between 1999 and 2001 (http://www.cdc.gov/ncidod/dvbid/westnile/qa/cases.htm).1 Illinois reported the highest numbers of human cases (884) and deaths (66) in 2002 with 100 counties reporting WNV-positive birds, mosquitoes, humans, or horses (http://www.idph.state.il.us/envhealth/wnv.htm).2 A similar high level of transmission activity occurred in 2003 in the Great Plains and WNV was reported from 2,289 counties in 46 states, but epidemic transmission in Illinois was markedly lower in that year, with 54 cases and one death and transmission activity reported from 77 counties.1,2

Culex species appear to be the predominant vectors in the enzootic and epizootic bird-mosquito transmission cycles of WNV and three large urban outbreaks in the 1990s in the eastern hemisphere (Romania and southern Russia) and western hemisphere (New York City) implicated Culex pipiens L for the first time as the primary vector.3 An important component for developing vector and disease management programs is an understanding of the population dynamics of the mosquito vectors.4 Since its introduction into North America, the vectors of WNV have proven to vary regionally, homologous to that observed for the closely related flavivirus St. Louis encephalitis virus (SLEV).5,6 Although other species are vector competent under laboratory conditions, the predominant WNV-positive species from field collections in the midwest and parts of the northeastern United States have been Cx. pipiens and Cx. restuans Theobald.68

The majority of overwintering Cx. pipiens and Cx. restuans are mated, nonblood-fed females.9,10 Diapause termination in Cx. pipiens is dependent upon juvenile hormone biosynthesis11; however, in most temperate areas, the initiation of blood feeding in the spring is dependent upon temperature.12 The time from egg raft oviposition to adult emergence generally takes from 8 to 12 days in east-central Illinois13 and varies depending mainly on water temperature, nutrient quantity and quality, and larval crowding.12 Culex restuans egg rafts are often first detected between mid-April and May in central states such as Illinois, Iowa, and Indiana, and it rapidly becomes the dominant Culex species until June or early July. In contrast, Cx. pipiens is typically a rare species in the midwest from April to June and oviposition peaks between August and early September.1316 Crossover of the two Culex species is defined as the time during this transition when the relative proportions of the two species are equal. The shift in abundance from an early season dominance of Cx. restuans to a late season dominance of Cx. pipiens has led to the suggestion that the former species may initiate enzootic cycles, whereas the latter species may amplify the number of infected avian hosts.17,18 In addition to amplifying WNV among birds, Cx. pipiens is also considered the main epidemic vector in the midwest of WNV and SLEV. This is based on a demonstrated vector competency in the laboratory and a seasonal pattern of infection and abundance that both precedes and parallels the temporal pattern of human cases for the two flaviviruses.5,7,8,1922

Although the general pattern of crossover is similar over a broad geographic area, the exact timing of when the two species are in equal abundance is quite variable (ranging between July and September in central Illinois) and probably represents an interaction of meteorologic, ecologic, and density-dependent factors.14,23 A previous study indicated that annual shifts in crossover were not related to any obvious differences in environmental conditions between years.14 The aim of our study was to discover to what extent meteorologic and climatic factors play a role in the relative abundance of Cx. pipiens. This report addresses this issue and whether crossover can be predicted based on known meteorologic conditions. Our long-term goal is to develop a climatic index that accurately reflects the temporal abundance of this potential amplification vector of WNV and SLEV to be used in flavivirus risk models. Unfortunately, there is no known data set for the absolute abundance of Cx. restuans and Cx. pipiens in an area of WNV transmission; however, the relative proportion of the two species has been recorded in the Urbana-Champaign, Illinois area for more than a decade using oviposition traps.

MATERIALS AND METHODS

Since 1988, Culex oviposition has been monitored in the Urbana-Champaign area (40°6′ N, 88°15′ W) of east-central Illinois by the Illinois Natural History Survey. Relative abundance of Cx. pipiens and Cx. restuans were estimated by collecting egg rafts daily throughout the two cities in oviposition buckets baited with a rabbit chow infusion.15 The emerging larvae were identified to species in the third or fourth instar. Proportions of the two species were computed and plotted to determine the date of crossover, which is the date when the number of egg rafts from the two species are equal in pooled data from the various collection sites.

Observations of daily maximum temperature (Tmax), daily minimum temperature(Tmin), and daily precipitation were collected by the Illinois State Water Survey at its headquarters located in Champaign on the University of Illinois campus. This is an official cooperative observer site of the National Weather Service (site name: Urbana, site identification number: 118740). Temperature data were not available for each mosquito collection site. The Urbana site was the nearest climate station within the city limits (within 1–7 km of the study sites). The climate station is located on the southern edge of Champaign-Urbana (population approximately 100,000) in a grassy area surrounded by a mixture of buildings, trees, roads, and agricultural fields. All of the mosquito collection sites were in residential neighborhoods with land cover of similar character to the climate station site. Because of the lack of topographic variability in this area, the close proximity to study sites, and the similarity in land cover between the climate station and the collection sites, we concluded that the meteorologic observations from the climate station provided reasonable estimates for all the collection sites.

A cursory comparison of temperature data with the crossover dates indicated that warmer (cooler) summers were characterized by earlier (later) crossover dates. Many studies have found that one measure of the thermal environment, degree days (DDs), is related to insect development.24 Degree days are calculated as follows:

DD={TmeanTbase if Tmean>Tbase0 if TmeanTbase

where DD = number of degree days for a particular day, Tmean= mean temperature of the day, and Tbase= base temperature. The most appropriate value of Tbase varies by species. In the present study, DDs with different Tbase were investigated to determine whether these were related to the crossover date. In addition, two other measures of the thermal environment were investigated: number of days that Tmax was greater than an upper threshold temperature and number of days Tmin was below a threshold temperature. The deviation of Tmax and Tmin above or below maximum and minimum thresholds, respectively, are termed exceedances in this report. As with DDs, a series of upper and lower thresholds were correlated to crossover dates. Although these latter two temperature threshold exceedances are not commonly used to reflect insect development, our preliminary inspection of the data suggested a possible relationship. In addition, the analysis was performed for two periods: 1) last spring freeze to crossover, and 2) January 1 to crossover.

A simple linear function of the following form was assumed to express the dependence of the time of crossover on the thermal measures:

Dcross=SΔT+I

where Dcross = day of year of crossover, ΔT = thermal measure (degree day accumulation or number of days with threshold exceedances) accumulated from the beginning day to Dcross and expressed as a deviation from the climatic average, S = slope, and I = intercept. The thermal measure, ΔT, is calculated as follows:

ΔT=i=nN(tiai)

where ti = value of thermal measure on day i, ai = average value of the thermal measure on day i, n = beginning day for accumulation of the thermal measure (either 1 for January 1 or the day of year of the first spring freeze), and N = day of year of crossover. ΔTs were calculated with various values specified for DD base temperature or for the exceedance threshold temperatures. The average value of ai for the thermal measure was calculated for the period 1971–2000 using historical data from the same climate station. In this function (equation 2), the intercept physically represents the estimated crossover date for average temperature conditions and the slope represents the sensitivity of the crossover date to the thermal measure; an example is given in the next section. Hereafter, the term model will refer to equation 2 with one of the thermal measures specified for ΔT.

The three models were tested using a jackknife regression. For each year with historically observed crossover dates, a regression was run, using observed values of Dcross and ΔT, excluding the year to be modeled, to obtain values of I and S to be applied to the year excluded from the regression. Since the intent of the model is as an operational/forecasting tool, it was tested in a manner similar to the way it would be applied. That is, let j = current day of year. For each day, an estimated crossover day (Ecross) is calculated from equations 2 and 3:

Ecross=I+Si=nj(tiai)

If Ecross > j, then we assume that the crossover date has not yet occurred. On the first day that Ecross ≤ j, crossover is assumed to occur and this day is denoted as Ncross. Correlation coefficients between Ncross and observed values of Dcross for all the years with observations were calculated for each combination of thermal measure, base value (DD base or temperature threshold), and starting date. These correlations coefficients were used to identify the best predictive model.

RESULTS

Table 1 lists the dates when the first egg rafts of Cx. pipiens were detected and when crossover occurred (equal number of egg rafts for both Culex species). The mean number of days to crossover from January 1, based on 13 years of data for 1988–2003 was 219 days (SD = ± 22 days). The crossover day of year varied from 191 to 255, a range of approximately 9 weeks.

A scattergram of crossover day of year versus the number of days with Tmax exceeding 27°C from January 1 to the day of crossover (Figure 1) illustrates the inverse relationship between the thermal environment and the day of crossover. The relationship appears to be linear, a characteristic common to all of the following combinations of model type and base/threshold temperature. The slope of the linear fit is −1.4 days/day and the intercept is 219 days. Thus, for an average number of days with Tmax exceeding 27°C, the crossover day of year is expected to be 219 (August 7). For every additional (compared with the average) day above 27°C, the crossover day of year on average is earlier by 1.4 days.

Correlation coefficients were calculated for the model using DDs as the thermal measure for base temperatures from 5°C to 25°C (Figure 2). For the case of January 1 as the starting date for accumulation (solid curve), the r2 values are between 0.6 and 0.7 for base temperatures between 5°C to 22°C, with the maximum value of 0.68 occurring at 17°C. For the case of the last freeze as the starting date for accumulation (dashed curve), the shape of the correlation curve is similar to that for a January 1 starting date. The magnitudes of the correlations are lower, with the highest r2 value being 0.54 for a base temperature of 11°C.

Correlation coefficients for the number of Tmax exceedances as a function of a threshold temperature (Figure 3) show a pronounced dependence on threshold. For the case of January 1 as the starting date for accumulations (solid curve), maximum r2 values > 0.64 occur for Tmax thresholds of 27°C and 28°C. For the case of the last freeze as the starting date (dashed curve), the maximum r2 value is 0.54 at thresholds of 27°C and 28°C.

The correlation coefficients for number of Tmin exceedances as a function of the threshold (Figure 4) show that as with Tmin exceedances, the r2 values exhibit substantial dependence on temperature thresholds. For the case of January 1 as the starting date (solid curve), there is a peak r2 value of 0.52 at 20°C. For the case of the last freeze as the starting date (dashed curve), the maximum r2 value is only 0.27 at a threshold of 20°C. Thus the best correlation coefficients for deviations below Tmin at either starting date were less than those for correlations of deviations above Tmax.

DISCUSSION

West Nile virus by its very nature of having visible signs of transmission (dead birds), recurrent outbreaks, and high infection rates in vectors provides a unique opportunity to investigate the intricacies of the transmission cycle and the impact of meteorologic and ecologic variables on it. In our study, we focused on modeling the impact of temperature on the timing of the Cx. pipiens increase in abundance, which coincides with the amplification phase of WNV transmission in Illinois over the past three years (Novak RJ, Lampman RL, Gu W, unpublished data). The DD and Tmax exceedance models explain more than 65% of the observed variance in crossover date for some values of the base/threshold temperature, whereas the best Tmin exceedance model explains up to 52% of the variance. Also, the use of January 1 as the starting date for the accumulation of the thermal measure generally results in higher values of the explained variance compared with the use of the last freeze as the starting date.

Using as examples those cases with the highest correlations, the DD base of 17°C and the Tmax threshold of 27°C were chosen to illustrate the behavior of two operational models applied to the historical data. The model estimates of the crossover date obtained from the jackknife regression (Figure 5) show that the large interannual variations in crossover date are predicted rather well. There is little difference between the two models. In one year (1996), the model estimates are 3–4 weeks early. In half of the years, the model estimates are within one week of observed crossover. In other years, the model estimates are within 1–2 weeks of observed. Thus, the DD and Tmax exceedance models perform with approximately equal accuracy when using optimum values of the base/threshold temperatures. Surprisingly, the accuracy of the DD model was relatively insensitive to base temperature, whereas the Tmax exceedance model was quite sensitive to the temperature threshold chosen.

Since the use of a fixed calendar date (January 1) produces higher correlations than the use of last freeze date (which varies from year to year), these models are very straightforward to apply. A remaining question in terms of a climate index is its applicability to other regions of the country.

The initial evidence from field studies suggests that WNV is similar to SLEV in that the main enzootic and epizootic vectors may vary regionally in the United States, but primarily involve Culex species.18,25 Although considerable emphasis has been placed on the potential role of many mammal-feeding mosquito species as bridge vectors, particularly Stegomyia albopicta and Ochlerotatus japonicus,26 there is little field evidence that supports these non-Culex species as important epidemic vectors in the east-central United States.27 In Illinois, S. albopicta is common throughout the southern quarter of the state, but is only sparsely distributed in the central and northern regions. Neither of these species were present in Champaign-Urbana. The only species with a significant number of positive pools from dry ice-baited traps and gravid traps were Culex species. The most common non-Culex species in which WNV RNA was detected in central Illinois was Aedes vexans, although the number of pools and infection rates were well below 1 per 1,000 mosquitoes (Novak RJ, Lampman RL, Gu W, unpublished data).

Major outbreaks of SLEV and WNV tend to occur when temperatures are above average and rainfall is average to below average, although outbreaks may be preceded by above average rainfall.4,2830 High temperatures may favor flavivirus transmission by increasing the development rate of the vector, decreasing the interval between blood meals, and increasing the virus replication rate and magnitude of infection in mosquitoes, thus considerably shortening the extrinsic incubation period.28,31 For SLEV, temperature exceedances and DD accumulations above specific temperature limits have been reported to correlate with detection of the flavivirus or with transmission rates.28,32,33 Our study suggests that temperature plays an additional role in transmission of flaviviruses by altering the crossover date of Cx. restuans and Cx. pipiens. In other words, assuming Cx. pipiens is the major amplification and potential bridge vector to mammals in Illinois,34 its earlier appearance could provide a longer period for transmission to build to a threshold where incidental hosts become involved. Meteorologic factors or indices (e.g., drought or water table depth) have been suggested as important for bringing vectors in close association with avian hosts.4

The model for crossover of Cx. restuans and Cx. pipiens presented here is probably relevant for much of the east-central United States; however, latitudinal variation in crossover is poorly defined, thus making it difficult to test this hypothesis from the literature. Culex restuans oviposition activity in Illinois, Iowa, and Indiana begins between mid-April and May, and it becomes the dominant Culex species until June or early July. As the number of Cx. restuans decrease, Cx. pipiens oviposition increases, peaking between August and early September.1316 In southern Canada, the temporal pattern of abundance is similar; Cx. restuans is most abundant during the spring and early summer, whereas Cx. pipiens reaches its peak abundance in late summer to early fall.35,36

An experimental application of these models as an operational/forecasting mode was performed using temperature data for 2004. The approach used to test the models (see equations 4 and accompanying discussion) can provide an estimate of the crossover date on any day prior to the crossover. Implicit in such an estimate is that the remainder of the year, up to the crossover day, can be characterized by average temperature conditions, although additional information can be provided using historical climate data.37 Specifically, we assumed that the past climate history provides an envelope of what may happen during the rest of the current year. Furthermore, we assumed that the probability of what will happen during the rest of the current year is equal to the past frequency of conditions. Each year in the historical climate database is assumed to be one scenario for the outcome of the remainder of the year. To apply this concept, the temperature time series for one scenario was assumed to be the combination of the actual observed data for 2004 up to the current date plus the observed temperature data from some past year for all days after the current date. An estimated crossover date was calculated using equation 4 stepping forward from the current date until Ecross ≤ j. This process was repeated 104 times using each year from 1900 to 2003 as a possible scenario for the remainder of the year. These 104 values were sorted from earliest to latest day of year. The result is a probability distribution of crossover dates; thus, this provides an estimate of both the variance and the mean of estimated crossover date. The earliest, latest, median (average of rank 52 and 53 values), the 90% probability of exceedance (rank 10), and the 10% probability of exceedance (rank 94) values were extracted and plotted (Figure 6).

Probability distributions were generated for each week of 2004 beginning with May 1 (Figure 6). For estimates made during May, the range of estimated crossover days of year is large (187–250), reflecting the fact that Cx. pipiens population buildup is in its early stages and the future weather is the determining factor. As the season progresses and the weather for an increasing fraction of the year has been realized, the range of the thermal measure narrows. Or, in a more physical sense, the mosquito population dynamics has progressed closer to crossover and the uncertainty about its timing has lessened. Eventually, all of the curves converge to a single value (236), the final estimated value of crossover for 2004. Another interesting feature is the upward trend in the median and other curves (Figure 6). This is a result of weather conditions during the summer of 2004, which was the fifth coolest since 1889. Because of the cool temperatures, the probability distribution shifted to later crossover dates as the summer progressed. The final estimated value (236) was within the range of the initial (May 1) probability distribution, but on the upper end of the range. The information presented in Figure 6 provides insights into the uncertainties in crossover date as the season progresses.

Integrated mosquito management typically focuses on larval control; however, when this is inadequate, adult control measures may become necessary.38 Our model provides a method for the timing of adult control near the crossover date, which our research has shown precedes the peak infection rates in Culex (Novak RJ. Lampman RL, Gu W, unpublished data).

Table 1

Last spring freeze data and first appearance and crossover date for Culex pipiens in Urbana-Champaign, Illinois

YearLast spring freezeFirst Cx. pipiens in trapCrossover date (day of year)
1988April 20June 27July 10 (192)
1990April 18June 19August 7 (219)
1991April 11June 12July 10 (191)
1992April 28July 10September 11 (255)
1994April 8June 15July 25 (206)
1995April 5June 14July 29 (210)
1996April 27July 1September 9 (253)
1997April 18June 24August 31 (243)
1998March 24June 10July 20 (201)
1999March 28June 13July 29 (210)
2001April 18June 17August 22 (234)
2002April 7June 11July 28 (209)
2003April 13June 14August 23 (235)
Figure 1.
Figure 1.

Deviation from the long-term average of the number of days with a daily maximum (max) temperature approximately 27°C versus the day of year of crossover from Culex restuans to Cx. pipiens. The number of days is accumulated from January 1.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

Figure 2.
Figure 2.

Square of the correlation coefficient between crossover date and accumulated degree days as a function of degree day base for data from the period 1988–2003 (except for 1989, 1993, and 2000). Degree days are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

Figure 3.
Figure 3.

Square of the correlation coefficient between cross-over date and accumulated number of maximum temperature (Tmax) exceedances as a function of exceedance threshold for data from the period 1988–2003 (except for 1989, 1993, and 2000). Threshold exceedances are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

Figure 4.
Figure 4.

Square of the correlation coefficient between cross- over date and accumulated number of minimum temperature (Tmin) exceedances as a function of exceedance threshold for data from the period 1988–2003 (except for 1989, 1993, and 2000). Threshold exceedances are accumulated from January 1 (solid line) or the date of the last spring freeze (Frz) (dashed line).

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

Figure 5.
Figure 5.

Time series of the day of year of crossover from Culex restuans to Cx. pipiens for 1988–2003 for observed (Obs) data (thick solid line), estimates from a degree day model (DD Mod) using base temperature of 17°C (dashed line), and estimates from a maximum temperature (Tmax) exceedances model (Tmax Mod) using a thresh-old of 27°C (thin solid line).

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

Figure 6.
Figure 6.

Probability distribution of estimated crossover date as a function of the date on which the estimation was made. The probability distribution was computed from 104 scenarios, each scenario representing one year from the period 1900–2003. Five values of the probability distribution are displayed: earliest, 90% probability of exceedance, median, 10% probability of exceedance, and latest.

Citation: The American Journal of Tropical Medicine and Hygiene Am J Trop Med Hyg 74, 1; 10.4269/ajtmh.2006.74.168

*

Address correspondence to Robert J. Novak, Medical Entomology Program, Illinois Natural History Survey, Champaign, IL 61820. E-mail: rjnovak@uiuc.edu

Authors’ addresses: Kenneth E. Kunkel, Illinois State Water Survey, Champaign, IL 61820, E-mail: kunkel@hercules.sws.uiuc.edu. Robert J. Novak, Richard L. Lampman, and Weidong Gu, Medical Entomology Program, Illinois Natural History Survey, Champaign, IL 61820, E-mails: rjnovak@uiuc.edu, rlampman@inhs.uiuc.edu, and wgu@inhs.uiuc.edu.

Financial support: This work was supported by the National Oceanic and Atmospheric Administration (NOAA) under contract number EA133E-02-CN-0027, Centers for Disease Control and Prevention grant PHS U50/CCU52051, and the Illinois Waste Tire Fund, Illinois Department of Natural Resources.

Disclaimer: Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of NOAA or the Illinois State Water Survey.

REFERENCES

  • 1

    Centers for Disease Control, 2004. West Nile Virus. Cited January 20, 2005. Available from http://www.cdc.gov/ncidod/dvbid/westnile/qa/cases.htm

  • 2

    Illinois Department of Public Health, 2005. West Nile Virus. 2003. Cited January 20, 2005. Available from http://www.idph.state.il.us/envhealth/wnv.htm

  • 3

    Hayes CG, 2001. West Nile virus: Uganda, 1937, to New York City, 1999. Ann N Y Acad Sci 951 :25–37.

  • 4

    Day JF, 2001. Predicting St. Louis encephalitis virus epidemics: lessons from recent, and not so recent, outbreaks. Annu Rev Entomol 46 :111–138.

    • Search Google Scholar
    • Export Citation
  • 5

    Tsai TF, Mitchell CJ, 1989. St. Louis encephalitis. Monath TP, ed. The Arboviruses: Epidemiology and Ecology. Boca Raton, FL: CRC Press, 113–143.

  • 6

    Bernard K, Kramer L, 2001. West Nile virus activity in the United States, 2001. Viral Immunol 14 :319–338.

  • 7

    Turrell MJ, O’Guinn M, Oliver J, 2000. Potential for New York mosquitoes to transmit West Nile virus. Am J Trop Med Hyg 62 :413–414.

  • 8

    Turrell MJ, O’Guinn M, Dohm DJ, Jones JW, 2001. Vector competence of North American mosquitoes (Diptera: Culicidae) for West Nile virus. J Med Entomol 38 :130–134.

    • Search Google Scholar
    • Export Citation
  • 9

    Bailey CL, Faran ME, Gargan TP II, Hayes DE, 1982. Winter survival of blood fed and non-blood-fed Culex pipiens. Am J Trop Med Hyg 31 :1054–1061.

    • Search Google Scholar
    • Export Citation
  • 10

    Mitchell CJ, 1988. Occurrence, biology, and physiology of diapause in overwintering mosquitoes. Monath TP, ed. The Arboviruses: Epidemiology and Ecology. Volume I. Boca Raton, FL: CRC Press, 191–217.

  • 11

    Redio J, Chen M-H, Meola R, 1999. Juvenile hormone biosynthesis in diapausing and nondiapausing Culex pipiens (Diptera: Culicidae). J Med Entomol 36 :355–360.

    • Search Google Scholar
    • Export Citation
  • 12

    Vinogradova EB, 2000. Characteristics of natural populations of Culex p. pipiens in Russia and the neighbouring countries. Golovatch SI, ed. Culex pipiens pipiens Mosquitoes: Taxonomy, Distribution, Ecology, Physiology, Genetics, Applied Importance and Control. Sofia, Bulgaria: Pensoft Publishers, 116–149.

  • 13

    Lampman RL, Novak RJ, 1996. Oviposition preference of Culex pipiens and Culex restuans for infusion-baited traps. J Am Mosq Control Assoc 12 :23–32.

    • Search Google Scholar
    • Export Citation
  • 14

    Coraso BG, Munstermann LE, 1984. Identification by electrophoresis of Culex adults (Diptera:Culcidae) in light-trap samples. J Med Entomol 21 :648–655.

    • Search Google Scholar
    • Export Citation
  • 15

    Lee JH, Rowley WA, 2000. The abundance and seasonal distribution of Culex mosquitoes in Iowa during 1995–97. J Am Mosq Control Assoc 16 :275–278.

    • Search Google Scholar
    • Export Citation
  • 16

    Lampman R, Hanson, Novak R, 1997. Seasonal abundance and distribution of mosquitoes at a rural waste tire site in Illinois. J Am Mosq Control Assoc 13 :193–200.

    • Search Google Scholar
    • Export Citation
  • 17

    Luby JP, Sulkin SE, Sanford JP, 1969. The epidemiology of St. Louis encephalitis: a review. Annu Rev Med 20 :329–350.

  • 18

    Mitchell CJ, Francy DP, Monath TP, 1980. Arthropod vectors. Monath TP, ed. St. Louis Encephalitis. Washington, DC: American Public Health Association, 313–379.

  • 19

    Helson BV, Surgeoner GA, Wright RE, 1980. The seasonal distribution and species composition of mosquitoes (Diptera: Culicidae) collected during a St. Louis Encephalitis Surveillance Program from 1976–1978 in southwestern Ontario, Canada. Can Entomol 112 :865–874.

    • Search Google Scholar
    • Export Citation
  • 20

    Monath TP, 1980. Epidemiology. Monath TP, ed. St. Louis encephalitis. Washington, DC: American Public Health Association, 239–312.

  • 21

    Campbell GL, Marfin AA, Lanciotti RS, Gubler DJ, 2002. West Nile virus. Lancet Infect Dis 2 :519–529.

  • 22

    Gubler DJ, Campbell GL, Nasci R, Komar N, Petersen L, Roehrig JT, 2000. West Nile virus in the United States: guidelines for detection, prevention, and control. Viral Immunol 13 :469–475.

    • Search Google Scholar
    • Export Citation
  • 23

    Madder DJ, Surgeoner GA, Helson BV, 1983. Number of generations, egg production, and developmental time of Culex pipiens and Culex restuans (Diptera: Culicidae) in southern Ontario (Canada). J Med Entomol 20 :275–287.

    • Search Google Scholar
    • Export Citation
  • 24

    Baker DG, Sharratt BS, Chiang HC, Zandlo JA, Ruschy DL, 1984. Base temperature selection for the prediction of European corn borer instars by the growing degree-day method. Agricult Forestr Meteorol 32 :55–60.

    • Search Google Scholar
    • Export Citation
  • 25

    Reisen W, Lothrop H, Chiles R, Madon M, Cossen C, Woods L, Husted S, Kramer V, Edman J, 2004. West Nile virus in California. Emerg Infect Dis 10 :1369–1378.

    • Search Google Scholar
    • Export Citation
  • 26

    Turell MJ, Dohm DJ, Sardelis MR, O Guinn ML, Andreadis TG, Blow JA, 2005. An update on the potential of North American mosquitoes (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol 42 :57–62.

    • Search Google Scholar
    • Export Citation
  • 27

    Kilpatrick AM, Kramer LD, Campbell SR, Alleyne EO, Dobson AP, Daszak P, 2005. West Nile virus risk assessment and the bridge vector paradigm. Emerg Infect Dis 11 :425–429.

    • Search Google Scholar
    • Export Citation
  • 28

    Monath TP, Tsai TF, 1987. St. Louis encephalitis: lessons from the last decade. Am J Trop Med Hyg 37 (Suppl 3):40S–59S.

  • 29

    Hubalek Z, 2000. European experience with the West Nile virus ecology and epidemiology: could it be relevant for the new world? Viral Immunol 13 :415–426.

    • Search Google Scholar
    • Export Citation
  • 30

    Ward MP, Levy M, Thacker HL, Ash M, Norman SKL, Moore GE, Webb PW, 2004. Investigation of an outbreak of encephalomyelitis caused by West Nile virus in 136 horses. J Am Vet Med Assoc 225 :84–89.

    • Search Google Scholar
    • Export Citation
  • 31

    Dohm DJ, O’Guinn ML, Turell MJ, 2002. Effect of environmental temperature on the ability of Culex pipiens (Diptera: Culicidae) to transmit West Nile virus. J Med Entomol 39 :221–225.

    • Search Google Scholar
    • Export Citation
  • 32

    Reeves WC, Hammon WM, Longshore WA Jr, McClure HE, Geid AF, 1962. Epidemiology of the Arthropod-Borne Viral Encephalitides in Kern County, California, 1943–1952, Volume 4. Public Health, Berkeley, CA: University of California.

  • 33

    Hess AD, Cherubin CE, LaMotte LC, 1963. Relation of temperature to activity of western and St. Louis encephalitis viruses. Am J Trop Med Hyg 12 :657.

    • Search Google Scholar
    • Export Citation
  • 34

    Apperson CS, Harrison BA, Unnasch TR, Hassan HK, Irby WS, Savage HM, Aspen SE, Watson DW, Rueda LM, Engber BR, Nasci RS, 2002. Host-feeding habits of Culex and other mosquitoes (Diptera: Culicidae) in the Borough of Queens in New York City, with characters and techniques for identification of Culex mosquitoes. J Med Entomol 39 :777–785.

    • Search Google Scholar
    • Export Citation
  • 35

    Madder DJ, Macdonald RS, Surgeoner GA, Helson BV, 1980. The use of oviposition activity to monitor populations of Culex pipiens and Culex restuans (Diptera: Culicidae). Can Entomol 112 :1013–1018.

    • Search Google Scholar
    • Export Citation
  • 36

    Williams DD, Tavares CA, Kushiner DJ, Coleman JR, 1993. Colonization patterns and life-history dynamics of Culex mosquitoes in artificial ponds of different character. Can J Zool 71 :568–578.

    • Search Google Scholar
    • Export Citation
  • 37

    Duchon CE, 1986. Corn yield prediction using climatology. J Appl Meteor 25 :581–590.

  • 38

    Novak RJ, Lampman RL, 2001. Public health pesticides. Handbook of Pesticide Toxicology. Volume 1. Pesticide Risk Characterization. San Diego, CA: Academic Press, 181–201.

Save