Abstract
Populations of Ixodes scapularis Say nymphs were surveyed at a Lyme disease-endemic area for 8 consecutive yr (1998–2005) to characterize annual changes in abundance. Precipitation and temperature were also monitored over the period 1998–2004 to determine their potential value as predictors of tick abundance. Although both parameters showed annual variation, no statistical differences in the annual abundance of I. scapularis nymphs were observed over the 8-yr period. Our results suggest that precipitation and temperature were not predictive of the abundance of I. scapularis nymphs.
Lyme disease continues to be the most commonly reported vector-borne disease in the United States (CDC 2004), and over the 6-yr period of 2000–2005 alone, New Jersey reported an average of 2,708 confirmed Lyme disease cases annually. Although surveillance data are useful in identifying annual trends in Lyme disease incidence, they are of limited value in predicting future transmission risk. Because Lyme disease represents such a significant proportion of all reportable diseases in New Jersey and elsewhere, the ability to predict transmission risk could be an important adjunct to any prevention program, theoretically allowing public health agencies to increase awareness efforts and better target tick management and control programs (Brownstein et al. 2003).
There seems to be a direct relationship between the abundance of infected Ixodes scapularis Say nymphs and Lyme disease incidence (Falco and Fish 1989, Schulze et al. 1991, Falco et al. 1995, Mather et al. 1996, Stafford et al. 1998). In turn, precipitation and temperature, either alone or in concert, seem to have significant effects on tick populations and/or Lyme disease incidence (McEnroe 1977, Jones and Kitron 2000, Subak 2003, McCabe and Bunnell 2004). In this study, we attempted to determine whether data obtainable from existing long-term weather recording stations could be used by a municipal-level health agency to predict local tick abundance, as a measure of disease transmission risk, by contrasting simple temperature and precipitation indices to the abundance of I. scapularis nymphs over an 8-yr period. Our working hypothesis was that year-to-year abundance of nymphs would vary in response to differences in annual precipitation and temperature patterns.
Materials and Methods
Study Area.
Naval Weapons Station Earle (NWS Earle) is a 41-km2 military facility situated within an historic Lyme disease–endemic area in central New Jersey (Bowen et al. 1984). I. scapularis has been shown to be consistently abundant in the study area (Schulze et al. 1986, 2006; Schulze and Jordan 1996). The forest canopy consisted of pitch pine (Pinus rigida Miller), white oak (Quercus alba L.), red oak (Q. rubra L.), and chestnut oak (Q. prinus L.). The understory comprised saplings and seedlings of the canopy species, American holly (Ilex opaca Aiton), and sassafras (Sassafras albidum [Nuttall] Nees) over a patchy, often dense thicket of highbush blueberry (Vaccinium corymbosum L.), lowbush blueberry (V. angustifolium Aiton), huckleberries (Gaylussacia spp.), sweet pepperbush (Clethra alnifolia L.), spicebush (Lindera benzoin L. Blume), laurels (Kalmia spp.), and greenbriar (Smilax rotundifolia L.).
Tick Collections.
We established 20, 100-m2 plots throughout NWS Earle in areas with a sparse shrub layer to facilitate collection of I. scapularis nymphs using drag sampling (Ginsberg and Ewing 1989, Schulze et al. 1997). Tick drags were constructed of a 1-m2 piece of light-colored corduroy fastened to a wooden dowel (1 cm) along the leading edge. Drags were worked along side of the investigator by means of a 2-m rope handle attached to the ends of the wooden dowel. During the 8-yr period 1998 through 2005, each plot was surveyed four times at approximately weekly intervals from mid-May to mid-June of each year, the peak activity period of I. scapularis nymphs in New Jersey (Schulze et al. 1986). Sampling was performed between 0800 and 1200 hours (Schulze and Jordan 2003) when vegetation was dry and wind speed was <10 km/h. Ticks adhering to drags were removed at 20-m intervals (Schulze and Jordan 2001), identified, counted, and returned to their respective plot. By sampling questing ticks throughout the seasonal activity peak, and during periods and under conditions when I. scapularis nymphs were most likely to quest, we attempted to avoid either under- or overestimating actual tick abundance (Schulze et al. 1997).
Weather Parameters.
All weather data were obtained from the Freehold, NJ, weather station operated by the South Jersey Resource and Development Council (Hammonton, NJ) located ≈6.6 km from NWS Earle. For the years 1998–2004, we grouped precipitation and temperature data by seasons, which roughly correspond to the various activity periods of I. scapularis nymphs (March-June) and larvae (July-September) in New Jersey (Schulze et al. 1986). Conditions encountered by larvae in the previous year (t - 1) would be expected to affect numbers of nymphs in the current year. We also calculated cumulative growing degree-days for the current nymphal season (1 April to 30 June) and for the larval season in the previous year (t - 1) (Ostfeld et al. 2006). Precipitation was expressed as cumulative daily precipitation (cm) for the same periods.
Statistical Analyses.
Analysis of variance (ANOVA) for repeated measures was used to determine the effects of sampling date and plot on numbers of questing ticks (Sokal and Rohlf 1981). Relationships between tick numbers and climate parameters were explored using Spearman correlation coefficients (Sokal and Rohlf 1981). All statistical tests were performed using Statistica analysis packages (StatSoft 2005).
Results
Tick Collections.
Four weekly sampling events conducted at 20 sites during the peak nymphal activity period for 8 consecutive yr (1998–2005) yielded 3,416 I. scapularis nymphs. The number of nymphs collected in any given year varied from 381 in 2000 to 633 in 2004. Tick abundance varied significantly between sample plots (F(19,480) = 3.62, P < 0.01) but not between years (F(7,480) = 0.90, P = 0.51). The mean of four weekly sampling events ranged between 4.7 ± 0.6 and 9.1 ± 3.9 (SE) nymphs/site, but these differences were not statistically significant (F(6,531) = 1.81; P = 0.09). However, there was considerable variability in individual weekly survey means within years, with the greatest (1.7–20.6 nymphs/site) observed in 1998. Nymphal abundance during the first week of sampling tended to be significantly lower than later in the season (F(3,531) = 4.99; P < 0.01).
Weather Parameters.
Total precipitation in spring and summer ranged from 19.9 cm (1999), the driest year of the period, to 33.5 cm (2001) the wettest growing season (Fig. 1). Both 2001 and 2004 were noticeably wetter in the early spring, and rainfall amounts remained well above the long-term monthly mean throughout both years, whereas there was a substantial drop in total precipitation from spring to summer months in 1998. Number of growing season degree-days >10°C varied from 1,412.5 (2001) to 1,638 (2002) (Fig. 1). The coolest years of the period were 2000 and 2003, when local temperatures were slow to warm into summer and remained substantially below the long-term monthly mean for the growing season. Nymphal abundance was not related to either precipitation or temperature in the current or previous years (Table 1).
Mean nymphal I. scapularis abundance (weekly mean ± SE) at NWS Earle 1998–2005 (a) contrasted with cumulative growing season precipitation (b) and cumulative growing season degree-days (c) recorded at Freehold, NJ, 1998–2005.
Mean nymphal I. scapularis abundance (weekly mean ± SE) at NWS Earle 1998–2005 (a) contrasted with cumulative growing season precipitation (b) and cumulative growing season degree-days (c) recorded at Freehold, NJ, 1998–2005.
Spearman rank correlation coefficients comparing I. scapularis abundance at NWS Earle and weather parameters, 1997–2005
Discussion
The idea that weather parameters can be used to predict tick abundance and disease incidence is predicated on the premise that I. scapularis requires relatively cool, high humidity conditions for survival and that wet springs and summers and milder winters favor tick survival, whereas drought conditions and colder winters lead to increased mortality (Subak 2003). However, we found that, although local precipitation and temperature varied between years during one or more tick activity periods, the abundance of I. scapularis nymphs showed little variation over 8 yr. Nymphal abundance was not related to either total precipitation or growing season temperature in the current or previous years, and these simple parameters were poor predictors of I. scapularis abundance. Ostfeld et al. (2006) also reported only weak support for precipitation- and temperature-related variables as effective predictors of variation in tick abundance and infection prevalence.
Our inability to detect a significant relationship between nymphal I. scapularis abundance and local weather conditions may reflect the fact that seasonal precipitation and/or temperature during the 8-yr period considered did not vary sufficiently to significantly affect tick abundance (Jones and Kitron 2000). It is more likely that tick abundance, and particularly that of a three-host tick with catholic host preferences such as I. scapularis, may not lend itself to direct prediction based on simple weather parameters. Unlike vector insects such as mosquitoes that are clearly dependent on a given habitat parameter (e.g., the availability of standing water breeding sites that can be predicted from rainfall patterns) (Woodruff et al. 2002), it is reasonable to expect that a particular tick generation, during a relatively complex 2-yr life cycle, may be exposed to a variable array of beneficial and adverse weather-related conditions that affect both it and its habitat. For example, meteorological factors directly influence the geographical distribution (McEnroe 1977, Schulze et al. 1984, Brownstein et al. 2003), survival (Daniels et al. 1989, Needham and Teel 1991, Stafford 1994, Lindsay et al. 1995, Bertrand and Wilson 1996), initiation of activity (Duffy and Campbell 1994, Clark 1995, Schulze et al. 2001), and questing behavior (Goddard 1992, Schulze et al. 2001, Schulze and Jordan 2003) of I. scapularis.
Prevailing weather patterns also indirectly impact tick abundance by affecting host populations and habitat (Smith et al. 1974, Vessey 1987, Mannelli et al. 1993, Jones and Kitron 2000). For example, drought conditions can alter small mammal host habitat, reducing host population density and indirectly tick numbers, and these relationships may also be habitat mediated (Mannelli et al. 1993). Such complex interactions of environmental factors may be more the rule than the exception. Bunnell (2006) found that abundance of I. scapularis was significantly correlated with the presence of certain soil types and that the relationship was significant in dry years but not in wetter years, suggesting interaction between ticks, edaphic conditions, and precipitation.
Killilea et al. (2008) pointed out that complex statistical relationships and models are difficult to interpret and often rely on unvalidated assumptions. For example, although long-term variations in weather patterns may result in changes in the overall abundance and distribution of I. scapularis, short-term variability in local weather is more likely to affect host-seeking behavior of ticks and the ability to accurately estimate tick populations based on isolated sampling events. Randolph and Storey (1999) showed that, during periods of drought, I. ricinus L. nymphs descend to moist lower vegetation layers for water replenishment and larvae avoid desiccation by becoming quiescent but that ticks seem to return to optimum questing patterns once microclimate conditions become favorable again. Therefore, weather conditions before a particular sampling event may have a profound influence on host-seeking activity and, thus, the ability to collect ticks, rather than accurately reflecting tick abundance at a specific site. Time of day and resultant fine-scale temporal variation in temperature and relative humidity may be more controlling of observed tick abundance by affecting efficiency of sampling, therefore significantly biasing population estimates for sympatric ixodid species and the assessment of tick-borne disease transmission risk (Schulze et al. 2001, Hubálek et al. 2003, Schulze and Jordan 2003).
Although advances in ecological research have enabled a more complete mapping of tick-borne disease risk that ostensibly enables health agencies to better target interventions (Brownstein et al. 2003, Jackson et al. 2006), there has been a marked lack of similar effort toward developing practical applications of research findings for the actual prevention of disease (Piesman and Eisen 2008). There may be several reasons for this. Given the substantial variation in apparent tick abundance within and between habitats, even in a given year (Schulze and Jordan 1996), and the great variation in weather patterns between climatic zones/physiographic provinces (Harnack et al. 2005), it is unlikely that a prediction of relative transmission risk generated at larger geographic scales (regional or state) would necessarily be accurate at the local level. Faced with shrinking health budgets (Stoto et al. 2005), it is also unlikely that local health agencies would have the resources necessary to rapidly assimilate, analyze, and exploit real-time tick abundance and weather data to predict transmission risk. Furthermore, it is unclear what an effective public health response would be to such information. At the local level, health agencies are generally limited to dissemination of education and awareness information designed to reduce exposure to infected ticks. However, although people may understand the risks of acquiring a tick-borne illness and the measures available to avoid it, they very often take no precautions despite personal experience or knowing an infected person (Hallman et al. 1995, Poland 2001, Hayes and Piesman 2003). Therefore, it is not clear that confounding the public health message with predictions on relative tick abundance and thus transmission risk for a particular year would be of much practical value. Unfortunately, until such time that reliable and easily interpreted predictors of I. scapularis nymph abundance and transmission risk are identified, the public health response may be confined to a consistent message about personal protective measures and tick reduction as the best means of reducing tick-borne disease incidence.
Acknowledgements
We thank T. Gentile and R. Green, NWS Earle, for support throughout the study. This work was supported by a Cooperative Agreement (U50/CCU206567-9) between the New Jersey Department of Health and Senior Services and the Centers for Disease Control and a subcontract agreement with the Yale University School of Medicine under USDA Cooperative Agreement 58-1265-7-067.
