This item is only available as the following downloads:
1 SPATIO TEMPORAL DYNAMICS OF AEDES TAENIORHYNCHUS MOSQUITO IN SARASOTA COUNTY FLORIDA By MIKE FALKNER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DE GREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2 2013 Mike Falkner
3 To my friends and family
4 ACKNOWLEDGMENTS First I would like to thank Dr. Liang Mao for his knowledge, support and patience throughout the length of this thesis. Second, I would like to thank my committee members and the rest of the UF geography department for their amazing contributions and guida nce during my career at UF Finally, I would like to thank Dr. Linthicum at the USDA for the data used in this project and his support over the years.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ............................ 4 LIST OF TABLES ................................ ................................ ................................ ...... 7 LIST OF FIGURES ................................ ................................ ................................ .... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .. 12 Background ................................ ................................ ................................ ....... 12 Study Objectives ................................ ................................ ............................... 13 Thesis Structure ................................ ................................ ................................ 13 2 LITERATURE REVIEW ................................ ................................ ........................ 14 Characteristics of Aedes taeniorhynchus ................................ .......................... 14 Studies on Mosquito Dispersal ................................ ................................ ......... 15 Mosquitoes and the Environment ................................ ................................ ..... 17 Effect of Wind on Mosquito Movement ................................ ............................. 18 Mosquito Surveillance and Control Programs ................................ ................... 18 3 STUDY AREA AND DATA COLLECTION ................................ ........................... 22 Study Area ................................ ................................ ................................ ........ 22 Mosquito Trap Data ................................ ................................ .......................... 22 Environmental Data ................................ ................................ .......................... 23 4 METHODOLOGY ................................ ................................ ................................ 27 Analysis of Spatio Temporal Dynamics of Aedes taeniorhynchus .................... 27 Population Weighted Centroid ................................ ................................ .......... 27 Inverse Distance weighting ................................ ................................ ............... 28 Exploring Effects of Environmental Factors on the Spatio Temporal Dynamics of Aedes taeniorhynchus ................................ ................................ ............... 29 Calculating Moving Distance and Direction of Population Centroids ................ 29 Wind Speed and Direction ................................ ................................ .......... 30 Wind Speed and Population Moving Distance ................................ ............ 32 Tide and mosquito population ................................ ................................ .... 33 Precipitation and Population size ................................ ............................... 33 5 RESULTS AND DISCUSSION ................................ ................................ ............. 36
6 Spatial temporal Dispersal of Aedes taeniorhyn chus ................................ ....... 36 Association between Wind Direction and Mosquito Moving Direction ............... 38 Association between Daily Wind Speed and Mosquito Dispersal Distance ...... 40 Association between Tide and Precipitation Levels and Mosquito Populations 41 Limitations and Future Research ................................ ................................ ...... 42 6 CONCLUSION S ................................ ................................ ................................ ... 61 LIST OF REFERENCES ................................ ................................ ......................... 63 BIOGRAPHIC AL SKETCH ................................ ................................ ...................... 68
7 LIST OF TABLES Table page 3 1 Description of data used in thesis ................................ ................................ 26 5 1 Hypothesis testing for R 2 statistics under the null hypothesis that R 2 =0 ...... 60
8 LIST OF FIGURES Figure page 2 1 Compo nents of a Mosquito Light Trap ................................ ........................ 21 3 1 Sarasota mosquito management zone locations with county location shown in green on Florida state map. ................................ ................................ ...... 24 3 2 Total monthly mosquito counts from 1992 1994 ................................ .......... 25 3 3 Locations for data sources used in the project.. ................................ .......... 25 4 1 Methods displayed as a flow chart ................................ ............................... 34 4 2 Examples of daily wind vectors created from hourly speed and directional data ................................ ................................ ................................ .............. 35 5 1 IDW maps showing Aedes taeniorhynchus density locations in relation to the population weighted centroid locations.. ................................ ....................... 45 5 2 IDW maps showing Aedes taeniorhynchus density locations in relation to the population weighted centroid locations. ................................ ....................... 46 5 3 IDW maps showing Aedes taeniorhynchus density locations in relation to the population weighted centroid locations. ................................ ....................... 47 5 4 IDW maps showing Aedes taeniorhynchus density locations in relation to the population weighted centroid locations. ................................ ....................... 48 5 5 Population weighted centroid map of Ae des taeniorhynchus for 1992. ....... 49 5 6 Population weighted centroid map of Ae des taeniorhynchus for 1993.. ...... 50 5 7 Population weighted centroid map of Ae des taeniorhynchus for 1994 ......... 51 5 8 Wind di rection in relation to mosquito moving direction ............................... 52 5 9 Wind direction in relation to mosquito moving direction ............................... 52 5 10 Mosquito eastern movement to wind direction ................................ ............ 53 5 11 Mosquito eastern movement to wind direction ................................ ............ 53 5 12 1992 1994 Wind s peed to distance between centroids ............................... 54 5 13 Above average wind speed to distance between centroids ......................... 54 5 14 Below average wind speed to distance between centroids ......................... 55
9 5 15 Eastern movement directions, wind speed to distance ................................ 55 5 16 Above average wind speed to east ern movement distance ........................ 56 5 17 Precipitation and tide to counts with 7 day lag shown on scatter plots, 5 18 Precipitation and tide to counts with 7 day lag shown on scatter plots, 1993 ................................ ................................ ................................ ............. 58 5 19 Precipitation and tide to counts with 7 day lag shown on scatter plots, 19 94 ................................ ................................ ................................ ............. 59
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements fo r the Degree of Master of Science SPATIO TEMPORAL DYNAMICS OF AEDES TAENIORHYNCHUS MOSQUITO IN SARASOTA COUNTY FLORIDA By Mike Falkner May 2013 Chair: Liang Mao Cochair: Andy Tatem Major: Geography Aedes taeniorhynchus is a potential mosqutio vector that can spreads important human and animal arboviruses in Florida, such as West Nile viruses St. Louis encephalitis, and Rift Valley fever To support the control and prevention of mosquito borne diseases, I explore d the spatio temporal dynamics of the Ae des t aen iorhynchus mosquito in Sarasota County, Florida with the aid of Geographic information technology Population weighted centroids from Sarasota mosquito management trap sites were calculated and used to show locations of density and movement across the stud y region. Distances and movements were calculated between sequential centroid locations and then compared with daily wind directions and speed. Daily tide and precipitation levels were compared to trapped Ae des t aeniorhynchus population numbers to analyze correlations between the two. Results show that t he highest population numbers for Aedes taeniorhynchus are consistent during the typical Florida summer months of elevated in May, June, July, and August. The map of p opulation w eighted c entroids displayed a marked dispersion to inland areas from the Gulf of Mexico coast over time. Surprisingly, t here are no
11 statistical correlations found between Ae des taeniorhynchus dispersal and daily wind speed, along with wind direction s in Sarasota County, Flo rida. However, there are stronger relationship s between Ae des t aeniorhynchus population size and tide levels, as well as precipitation levels With international travel on the rise and possible introduction of other arboviruses, Florida residents are facing an elevated risk of imported mosquito vectored illness. Knowledge of the temporal and spatial distribution of Ae des t aeniorhynchus would be critical for categorizing areas of varying risk for disease transmission. Since Aedes taeniorhynchu s is a coastal species with the capacity to transmit indigenous and exotic arboviruses and there is a potential for an introduction of exotic diseases into t he United States through shipping ports, enhanced surveillance and control measures need to be established in Florida My thesis is one such example of vector surveillance and show s the capability for understanding aerial movement of disease carrying vectors.
12 CHAPTER 1 INTRODUCTION Background Mosquito borne diseases involve the transmission of viruses and parasites from animal to animal, animal to person, or person to person, by the means of mosquito vectors These diseases are of great concern to the socio economy of the Florida State. S ince the introduction of West Nile virus (WNV) in 2001, a total of 309 infection cases and 19 fatalities have been reported in Florida [ 1 ] [ 2 ] Outbreaks of St. Louis Encephalitis virus (SLEV) have decreased in recent years but epidemics have still been documented in Florida for example, 223 human cases during 1990 [ 3 ] Eastern Equine Encephalitis (EEE) is another mosquito borne disease widely reported in Florida with an average of 70 human cases per year [ 4 ] With an increase of international travel and possible introduction of other arthropod borne viruses (arboviruses), Florida residents are facing an elevated risk of imported mosquito vectored illness [ 5 ] Aedes taeniorhynchu s, also known as the black salt marsh mosquito, is often associated with salt marshes along coastal areas in North, Central and South America including Florida [ 6 ] It s potential of vectoring human and animal arboviruses has been well documented [ 7 ] For instance, the SLEV Everglades, and WNV have all been isolated from Ae des taeniorhynchu s in Florida Recent studies in the lab oratory environment have found that it can also transmit epizootic strains of Venezuelan equine encephalomyelitis, eastern equine encephalitis, and Rift Valley fever viruses [ 7 ] Capable of vectoring multiple existing and emerging disease agents, this species of mosquitoes has brought potential risk to the public health of Florida.
13 To address this potential risk, Sarasota County Florida, has initiated a mosquito management program for disease surveillance and prevention, habitat elimination, larval management, and spraying. The effectiveness of this program is highly dependent on th e knowledge of how mosquito population s distribute over time and space including the Ae des taeniorhynchu s species The driving forces behind its distribution vary dramatically between locations and are complicate d by various factors, such as blood meal availability wind, rainfalls, flooding and human transportation etc. ,  For the Sarasota County l ittle is known about the spatio temporal dynamics of Ae des taeniorhynchu s population and movement The lack of s uch knowledge prevent s the county government from effectively evaluat ing risks of mosquito borne diseases and implementing control/surveillance strategies [ 8 ] S tudy Objectives T o fill this knowledge void, my objectives of this research have two folds 1) I attempt to describe the spatio temporal d ynamics of Ae des taeniorhynchus mosquito in Sarasota C ounty, Florida based on trap data from 1992 1994 and a GIS approach 2) I intend to explain the observed patterns using statistical analysis and tide, precipitation, and wind datasets. Thesis Structure The thesis is organized in six chapters. Following this introduction chapter, a literature review chapter introduces detailed information about the Ae des taeniorhynchus relevant studies on this species, and mosquito control methods. Chapter 3 describes the methodology for analyzing spatial temporal dynamics of Ae des taeniorhynchus population and movement The chapter that follows (C hapter 4 ) present and discuss analysis results The last chapter concludes this thesis and articulates the implications from findings.
14 CHAP TER 2 LITERATURE REVIEW Characteristics of Aedes taeniorhynchus Ae des taeniorhynchu s, also known as the black salt marsh mosquito, is normally associated in high numbers with salt marshes along coastal areas in North, Central and South America  They are characterized by bands of white scales across the upper sides of abdominal segments [ 9 ] It has the potential to be a critical vector of important human and animal arboviruses. Further, SLEV Everglades, and WNV have all been isolated from it in Florida, and the species can transmit epizootic strains of Venezuelan equine encephalomyelitis, eastern equine encephalitis, and Rift Valley fever viruses in the lab [ 7 ] The life cycle of Ae des taeniorhynchu s lasts about three weeks with some having the ability to live even longer [ 10 ] There are four stages in the life cycle, namely egg, larvae pupae, adult Female adult s of th is species often lay their eggs on dry ground where the eggs would emerge when flooded by rain or tide water [ 10 ] The egg placement is generally alon g a contour line at a specific elevation relative to the high water line and in depressions in the upper regions of salt marshes or mangrove swamps [ 9 ] It takes approximately 4 days for eggs to hatch into larve, the second stage in the life cycle. Even though the larvae can develop in any salinity from fresh to ocean water, breeding is known to take place only in regions along the coast [ 10 ] Mosquito larvae usually develop into adults in about 7 10 days [ 11 ] The newly emerged mosquitoes stay in the area where they developed into adults for a period of 12 24 hours. Over the next few days they will disperse over a large area [ 10 ]
15 Studies on Mosquito Dispersal Mosquito movement by flight is often referred to as the d ispersal or migration. Williams [ 12 ] uses the term dispersal to commonly describe random flights that are often a factor of wind direction, while migration is considered to be a purposeful flight linked to the biology and survival of the species. For my thesis I use the term dispersal as described by Service, et al. in the book Mosquito Ecology [ 13 ] Service et al. define d the dispersal as the full range of mosquito movement that covers a distance of just a few meters or many kilometers from a starting location. Mosquito dispersal can be goal orientated with a variety of objectives ranging from nectar feeding, blood feeding, m ating, or looking for a suitable location to lay eggs. Other dispersal behaviors seem to be more random because they seem to serve no special biological purpose and are mostly driven by wind Most of these wind driven flights are involuntary and are charac terized by a lack of control as to where the mosquito may end up [ 14 ] Examples of the long distance dispersal of mosquitoes can be seen amongst the species found in coastal salt marshes. One such case is with Ae des taeniorhynchus and the mass exodus es for which they are known [ 10 ] Aedes vig ilax has been discovered more than 60 miles from its coastal breeding sites and also on a boat 20 miles from the shore [ 15 ] In another instance, Aedes sollicitans have been caught 28 miles out at sea [ 16 ] and also 110 miles inland from the shore [ 17 ] Wind is not the only mechanism that drives mosquito movement. Human involvement from ships, trains, aircrafts, and other vehicles can also play a factor in species dispersal. A memorable example is the intro duction of Anopheles gambiae from Africa to Brazil via shipping route. The arrival of efficient malaria vector s caused the largest epidemic of malaria seen in the Americas infecting as many as 290,000 people
16 killing at least 26,000 [ 18 ] More recent examples involve the airport malaria cases from which infected mosq uitoes are transported to non malarious countries such as France, Switzerland, Belgium, England, Russia and the Netherlands [ 19 ] It is important to know the spatial and temporal patterns of mosquito species as a key factor in mosquito control and also disease surveillance. T he b iomarker approach has bee n widely used to understand the dispersal of mosquitoes , [ 16] The biomarker approach, also referred to as the mark recapture methods w as originally designed to estimate population size but is frequently used to study mosquit o dispersal, feeding behavior, and also survival rates. This type of study involves the marking of lab raised mosquitoes to be released into the wild for a recapture at alternate trap sites [ 14 ] A variety of methods and techniques for marking mosquitoes have been developed to evaluate movement including dyes, dusts, and radioactive isotopes [ 13 ] Based on the mark recapture data the population density ca n be plotted against the log of distance from the release point. A statistical model ( then can be built to fit the plotted trend, assuming the population decreases by a constant amount for equal multiples of the distance from the release po int. Here, is the number of insects caught at distance from the origin of dispersal, while and are constants [ 13 ] The alternative is to plot log population against distance. This model ( infers a decrease in population (y) by a constant proportion for each measure of distance from the release point (x) Although the bio marker approach is efficient to study mosquito population dynamics, it has several downfalls. The principle disadvantage of mark recapture methods is the n ecessity to recapture a relatively large proportion of the sample population in order to attain acceptable levels of accuracy. First, i t is known that
17 marking can affect mosquitoes so they do not disperse in their natural routines. The handling of mosquito es can affect them in such a way that they do not disperse far, or it can excite them and result in excessive dispersal [ 20 ] Second t he marking of a species can decrease its lifespan thus reducing numbers available for recapture o n successive days [ 14 ] The use of laboratory reared mosquitoes to be released in the wild m ay also produce adults that have abnormal survival rates and dispersal. The third issue is relevant to the dimensions of the study area. Past studies assume that marked mosquitoes are limited to the study area, which is often a short distance from the release point [ 14 ] T hese studies often ended up with very small recapture rates which lead to the question of what happen ed to the uncaught marked mosquitoes. It is not known if they are still locat ed within the study area and have evaded recapture, or completely emigrated from it all together leading to biases in subsequent analysis Mosquitoes and the Environment Numerous studies have investigated the role of environmental data on mosquito speci es [ 21 ] [ 22 ] [ 23 ] Many have focused on the role of weather and populations numbers in terms of disease spread. Takeda et al. [ 24 ] correlated changes in climate to multiple arboviruses and their vectors in Rhode Island. Ross River virus cases have been linked to ongoing winter and spring rainfall which in turn allows for a magnification of the virus in the abundant water sources [ 25 ] Tide and precipitation levels have been linked to Ae des taeniorhynchus population size [ 26 ] Ritchie et al. [ 11 ] has shown that summer tidal levels are a direct influence on the hatching of large broods of Ae des taeniorhynchus in Florida mangroves. Numerous studies have shown a relationship between disease vectors and precipitation [ 27 ] [ 28 ] [ 29 ] Research conducted in the state of Florida has also
18 public ized the role of high rain fall along with mosquito counts [ 30 ] [ 31 ] The rainy season in Florida spans from May October which covers the four month study period for the three years examined [ 32 ] Effect of Wind on Mosquito Movement There have been a wide rang e amount of studies published on the topic of wind driven insect movement. It is reported that f light behaviors among insects can rival the sophistication that has been seen in migrant birds and increase migration distance by 40% [ 33 ] Some insect species migrate only a short distance betwe en habitat locations but some are able to disperse on large scales that cover countries and continents [ 34 ] Chapman et al indicated that major wind borne migration of P. xylostella from t he Nether lands to southern England was responsible for the establishment of the U.K. population [ 35 ] Insect movement has been measured using different techniques ranging from Malaise traps [ 36 ] to airpla nes equipped with nets [ 37 ] to radar equipment [ 38 ] Downwind movement is not the only direction in which insect species will migrate. It has been published that the pollen beetle, Meligethes aeneus will migrate upwind towards oilseed rape fields [ 39 ] In a mark release recapture study carried out in New Hampshire on the tree hole mosquito Ochlerotatus triseriatus recaptures in the study were not related to the prevailing wind direction [ 40 ] Another mark release recapture study from Australi a involving Aedes aegypti which is in the same genus as my study species, indicates a statistically significant trend for upwind flight [ 41 ] Mosquito Surveillance and Control Programs Since the arriva l of early European settlers, mosquitoes have engaged in a th
19 century, outbreaks of yellow fever across the state took a tremendous toll on the population. In the city of Jacksonville alone, with a population of 26,800 during this time, the 1888 Yellow fever epidemic caused 5,000 people to fall ill resulting in 400 deaths and 10,000 people fleeing the city [ 42 ] These disease outbreaks led to the establishment of the Florida State Board of Health (FSBH) in 1889. The Florida Anti Mosquito Association (FAMA) was formed in 1922. Early control efforts focused on dewatering of ditches along with some dredge and fill work. The positive effects from these efforts led to the creation of state funds in 195 3 which allowed for permanent control work and the establishment of the Entomological Research Center in Vero Beach [ 42 ] Collecting adult mosquitoes can yield important information to a surveillance area. Increases in populations can be observed at specific trapping locations time. Once populations have been detected and identified, control measures such a s larviciding and adulticiding can be easily put in place. Knowing what species of mosquito is breeding can help vector control find the breeding source and take the correct control measures. After detection of population s, mosquitoes can also be tested fo r the presence of disease. The detection of a virus in a mosquito population that actively seeks human blood meals indicates a true potential for human disease outbreaks. Trap collections not only determine where control measures are needed, but also deter mine the effectiveness of control measures already in place [ 42 ] The New Jersey Light Trap (Figure 2 1) continues to be the most widely used adult mosquito trap in Florida. Adult mosquitoes are attracted to the all metal trap by a 25 watt light bulb attached beneath a wide conical top. Mosquitoes are attracted to the
20 light source then drawn into the trap by a downward blowing fan through a screened funnel. Once trapped by the fan the mosquitoes die within the killing jar containing an adulticide [ 42 ] In addition to mosquito trapping, p lacing sentinel chickens for an extended period across a variety of location is another method widely used for mosquito disease surveillance The chickens are bled once every two weeks year round except during the winter months. Blood samples are processed and tested and results are used to place proper control measures where viral activity is present [ 42 ] management methods. These methods make use of biological, cultural, physical, and chemical tools while emphasizing health and safety to the community. The county elimination, larval management, and spraying. The county focuses on prevention by advising developers and c ivil engineers to build and maintain storm water systems that do not create optimal mosquito habitat while habitat elimination puts the emphasis on controlling exotic plant specie s that support mosquito larvae. Sarasota County also regularly inspects for m osquito larvae and treat stagnant water that support growth. Spraying for adults is only done when necessary and is carried out corresponding to state guidelines. Mosquito management for Sarasota also works closely with other county health department offic es to monitor and prevent the spread of vector borne diseases [ 43 ]
21 Figure 2 1. Components of a Mosquito Light T rap [ 44 ] A) V ertical metal cylinder, B) conical roof, C) 25 watt light bulb attracts mosquitos, D) mesh screen to exclude larger insects, E) fan with electric motor, F) fine mesh funnel, G) Killing agent, H) insecticide strip inside jar, I) ventilated plastic cup to se parate insects from killing agent.
22 CHAPTER 3 STUDY AREA AND DATA COLLECTION Study Area miles south of Tampa Bay ( Figure 3 1 ). The county covers a total area of approximately 725 square miles with 37 miles of open shoreline along the Gulf of Mexico [ 45 ] A few prominent cities situated within the county are Sarasota, Venice, and North Port, as well as the town of Longboat Key. The county has a total population of about 381,000 permanent resid ents year round During the winter months the population size can reach to more than 450,000 due to visitors from up north heading down south for a warmer weather [ 45 ] Sarasota has a humid subtropical climate, with hot summers, mild winters, and high humidity year round. The rainy season lasts from June to September which Sarasota is lined with salt marshes along the coast. These coastal marshes are communities of vegetation in areas alternately swamped and drained by tide water. The n ame salt marsh summarizes conditions of the habitat and type of vegetation that encompass it. The salt marshes are filled with mostly grassy vegetation and can abundance o f salt marshes make it an ideal habitat for salt marsh breeding mosquito species. Mosquito Trap Data In order to analyze spatial temporal distribution of Aedes T aeniorhynchus population I have collected the mosquito trap data in Sarasota County Florida from
23 1992 1994. Mosquito trap data was created by the counties mosquito management program, and distributed by Center for Medical, Agricultural, and Veterinary Entomology (CMAVE) in Gainesville, Florida The data shows the count number of each mosquito species caught daily in light traps spanning the months between March to December and the years of 1992 94. Out of the 44 different species accounted for in the data Aedes taeniorhynchus was selected to be a focus of th is study due to its flight range and ability to vector dangerous arboviruses in the state. Sarasota County sets and collects the mosquito traps in 50 consistent locations throughout the county with the exception of the area covering the Myakka River state park ( F igure 3 1 ) Each trap location is found in a different area in the co unty referred to as a zone. There are 50 different zones in the county with each identified by a letter and number combination for example, R7 Each zone corresponds with one trap site located in the area. For this study only four months with the largest mosquito populations namely, May, June, July, and August, were used for analysis ( F igure 3 2 ) It is believed that these four months are the most representative to the spatial temporal distribution of Aedes t aeniorhynchus population Environmental Data To explore the driving forces of population dynamics amongst Aedes taeniorhynchus a variety of environmental data sets were collected, including the wind, tidal, and precipitation datasets of Sarasota County from 1992 1994 Their basic characteristics are summarized in Table 3 1. Specifically, the w ind data was obtained from the National Data Buoy Center under the National Oceanic and Atmospheric [ 46 ] The data recorded at station VENF1 located in Venice, Florida was used to represent Sarasota C ounty, which contain s information on hourly
24 wi nd speed and degree direction. Historical tide data was and currents historical data sets [ 46 ] Due to the data availability, only the data from S tation No. 8726520 about 40 miles north in St. Petersburg, Florida was used. The tidal data sets consisted of daily high and low values measured in feet from 1992 1994 For e (CLIMOD). The historical data sets were from S tation No. 089176 located in Venice, Florida and contained daily precipitation levels measured in inches. Locations for data sources are displayed below (Figure 3 3). Figure 3 1. Sarasota mosquit o management zone locations with county location shown in green on Florida state map.
25 Figure 3 2. Total monthly mosquito counts from 1992 1994 Figure 3 3. Locations for data sources used in the project S tation No. 089176 located in Venice, Florida. S tation No. 8726520 in St. Petersburg, Florida. 0 20000 40000 60000 80000 100000 2 3 4 5 6 7 8 9 10 11 12 Mosquito Count Month Monthly Totals of Aedes taeniorhynchus 1992 1994
26 Table 3 1. Description of data used in thesis Data Name Description Temporal Resolution Spatial resolution Source Mosquito Trap Data Daily mosquito counts Daily Trap zones USDA Wind Hourly wind speeds and directions Hourly Nearest weather station NOAA Tide Historical tide levels Daily Nearest tidal station NOAA Precipitation Precipitation levels Daily Nearest tidal station CLIM OD
27 CHAPTER 4 METHODOLOGY Analysis of Spatio Temporal D ynamics of Aedes taeniorhynchus To describe the spatial temporal dynamics of Ae des taeniorhynchus I examined the movement of population centroids and the population distributions over time. As shown in Figure 4 1, f or each observation day, t he weighted centroid method was employed to estimate the population centroid and meanwhile the inverse distance weighing (IDW) method was used to interpolate the population distribution. The details of both methods are described below. Population Weighted Centroid I used the Saraso ta mosquito management zone s to georeference mosquito trap data from 1992 to 1994. Each zone location which is represented by a polygon, within Sarasota County contains one corresponding mosquito trap site. Each trap stays in a fixed position within each zone unless forced to move due to county development or damage to a particular trap site S ince no specific latitude s and longi tude s were recorded for trap locations I calculated the centroid values of each zone in ESRI ArcGIS 10.0 as a possible substitute I then joined all relating mosquito trap information to the centroid p oints thu s giving values to all location s such as dates and counts of mosquitoes collected. To investigate mosquito dispersal patterns, I calculated a series of population weighted centroids for each collection date. By weighting the centroids by population I cou ld represent the population c entroid for an area rather than the geometric centroid. Population centroids are a useful tool in portraying the general trend of population movements [ 4 7 ] When modeling migration or dispersion it is assumed that population
28 weighted centroids are the most suitable out of the other methods available for calculating areal centroids [ 48 ] They are related to the distribution of the population within each area and can also be used for the calculation of migration distance [ 49 ] The population weighted mean cen troid locations of the mosquito population in Sarasota County can be found by multiplying the X and Y coordinates for each of the zones centroid population associated with that point. The mean of the weighted X coordinates and the means of the weighted Y coordinates define the location of the weighted mean center [ 50 ] The equations for the population weighted mean center are shown in Equation 4 1 and 4 1 In the two formulas above X and Y are the coordinates of the centroids within each zone, and w is the population size that is to be multiplied by the coordinates [ 50 ] Inverse Distance weighting To investigate if the po pulation weighted centroids are representative of actual distribution of populations a series of inverse distance weighted (IDW) maps was produced with contour lines showing varying mosquito densities. The IDW method has worked well in previous studies to interpolate mosquito densities across a landscape. Tachiiri et al [ 51 ] used the IDW in the creation of a raster based mosquito abundance map to evaluate West Nile v irus risk in British Colombia. Allen and Shellito [ 52 ] used IDW, among other methods, to characterize abundance patterns o f mosquito vectors of West Nile virus in Chesapeake, Virginia. Inverse distance interpolation is a weighted average of neighboring values. The weight given to each observation is inversely proportion to the distance between that
29 observations whereabouts and the initial grid point at which interpolation is sought after [ 53 ] The equatio n for IDW is shown in Equat ion 4 2 4 2 In this formula is the distance from the initial location to the th data location The coefficient is the weighting power and manages how fast the weights tend to zero as the distance increases, based on the assumption that observations are more similar the closer together they are. Exploring Effects of Environmental Factors on the Spatio Temporal D ynamics of Aedes taeniorhynchu s To explore the driving forces that changes the spatial temporal dynamics of Ae des taeniorhynchus I examined associations between 1) the daily displacement of population centroids and the wind speed, 2) the moving direction of population centroids and da ily prevailing wind direction, 3) the daily population size and precipitation, and 4) the daily population size and tidal height (Figure 4 1) The details of methodology design are described below. Calculati ng Moving Distance and Direction of P opulat ion C entroids I used the ArcGIS distance tool to measure the distance between the populat ion weighted centroid points T his tool help s draw a line or polygon on the map and obtain its length or area. The great circle d istance between points was given in meters for this study and can be calculated by using Equation 4 3 ) 4 3
30 where and are the latitudes, in degrees, of the respective coordinates and | is the absolute value of the difference of longitude between the respective coordinates [ 54 ] Secondly, the moving directions of population centroids between sequential centroid dates were estimated by Equation 4 4: 4 4 Because arctangent in excel returns the values of this equation in the range of 180 to +180, I had to normalize the results to a compass bearing in the range of 0 to 360. To do this I had to convert from the radian value given to degrees then utilize the modulo operatio n in excel. The modulo operation finds the remainder of division of one number by another. The syntax of the MOD (modulo) function is Equation 4 5 X =MOD( +360,360) [ 55 ] 4 5 The results of these angle computations were used for correlation analysis with wind data and will be discussed further in the following section of this chapter. The angle calculations were evaluated in Microsoft excel by using the latitude and longitude values at sequential population weighted centroid locations. Wind Speed and D irection To analyze how wind affect Aedes taeniorhynchus movement, I calculated the daily wind direction and average wind speed during t he study period, and attempted to relate them to the movement of mosquito population centroids With this information I sought to answer what role, if any, the wind speed and direction play in the movement of Aedes taeniorhynchus throughout Sarasota County Florida Based on the hourly
31 wind directions and speeds an hourly wind vector was produced for each mosquito observation date by plotting each hourly observation on the same line chart. The direction of the vector took the hourly wind direction, and the length of the vector was the hourly wind speed multiplied by an hour. A number of consecutive hourly wind vectors constitute a daily wind vector (Figure 4 2) B a sed on the daily wind vector, t he daily wind direction was calculated via the straight line fr om the value at point (the starting location) to (the ending location) on F igure 4 2 The equation in Microsoft excel used to calculate the angle was Equation 4 6. 4 6 Due to the default setting in used in excel to return values in radians I converted all outputs from the formula to degrees. The next step was to adjust the angle based on the quadrant by either subtracting 360 degree, or 90 degree, or 90 + degree, or finally 270 degree depending on what original quadrant the value was found. Subsequently I pairwise the wind directions to the moving directions of population centroid at the same observation date, and used correlation coefficient s to measure the asso ciation between them Since data for mosquito counts was not available every day, as was the case with the wind data, I designated two dates for the wind to be compared to the mosquito movement direction Because the mosquito movement direction was calcula ted between two collection dates, the two wind direction s selected were both between the two sequential count dates. The first wind direction for analysis was the day before the second collection date in each mosquito movement direction given this direction is the closest observation to the collection date.
32 The second was the day between the two collection dates that bore the highest average wind speed amongst all the days between the mosquito counts. This selection took into account the effects of the strongest winds on mosquito dispersal If by chance the day before the mosquito collection was the day with the highest average wind speed, then only one direction was used during that time frame. A series of scatter plots were produced to graphically depict the relationship between the wind directions and mosquito moving directions and the c orrelation coefficient s were estimated to indicate the strength of relationships. Wind Speed and Population Moving Distance Wind speed in relation to the moving distance of Aedes taeniorhynchus was also taken into consideration for this study. Similar as describe d in the above section, the wind speed selected was a date between the two sequential mos quito count dates. The day selected bore the highest avera ge wind speed amongst all the days between the mosquito counts. The wind speeds were then compared to the distance between each sequential mosquito centroid location to see if a higher wind speed equated to a longer distance between daily centroid location s. A series of scatter plots were produced to graphically depict the relationship between the wind speed s and mosquito moving distances and the c orrelation coefficient s were estimated to indicate the strength of relationships. I first compared all thr ee years, 1992 1994, to the date with the highest averaged wind speed. Next I explore d the relationship between extreme high wind values and moving distance. I define d an extreme high wind value as one standard deviation above the mean. I also investigated the association if any, of extreme low wind values with
33 moving distance. Extreme low wind values are defined as one standard deviation below the mean. Scatterplots were also produced to show eastern movements only and their relationshi p to wind speed. This was to remove all movement related to an angle on the negative x axis (on Figure 4 2) The positive x axis shows movements away from the coast. Anything west of the coast is located in the gulf of Mexico and was not sampled. Tide and mosquito population To examine the effect of tide on Aedes taeniorhynchus population, I used the maximum daily tide data in Sarasota County provided by NOAA Daily tide data, along with the mosquito data, from May to August of 1992 94 was assembled into a scatter plot. To asses tidal values with mosquito levels in this manner a time lag had to be applied to the mosquito data to account for the species incubation period (time between when eggs are laid and when they hatch) [ 56 ] Given previous studies conducted on Aedes taeniorhynchus during the Florida summer months a lag time of seven days was selected to weigh against the tidal values [ 57 ] After adjusting for the time lag a regre ssion line was calculated to show correlation between count numbers and tide events. Precipitation and Population size As was the case with the tidal data, the precipitation levels were plotted in a scatterplot and regressed against the time lagged mosquito counts. Once again the daily precipitation levels were analyzed with mosquito abundance numbers as well as the mean value.
34 Figure 4 1. Methods displayed as a flow chart
35 Fi gure 4 2 Examples of daily wind vectors created from hourly speed and directional data -100 -80 -60 -40 -20 0 20 40 60 80 100 -60 -40 -20 0 20 40 60 80 North and south distance in miles East and west distance in miles Daily Wind Vectors 1-May 6-Jun 9-Jun 2-Jun
36 C HAPTER 5 RESULTS AND DISCUSSI ON Spatial temporal Dispersal of Aedes taeniorhynchus From the maps produced I can see that the population weighted centroids are an accurate reflection of actual mosquito densities across Sarasota Counties. The first exampl e of this can be seen in the IDW map from June 10, 1992 (Figure 5 1 A ). From the date on this map you can see that the largest density of Aedes taeniorhynchus is located in the south western portion of the county along the coast. The population weighted me an center is very close to the actual location containing the highest mosquito density. Four days later on the 14 th of June you can see that the density levels have increased in the eastern portion of the map while decreasi ng on the west coast (Figure 5 1 B ). This result is suggestive of a decrease in the population due to an eastern dispersal from the coast. On June 17 th the population density is highest a gain along the coast (Figure 5 1 C ). Moving forward in time to the 21 st of June you can see the total population numbers decreasing by the density is continuing the eastern dispersal trend (Figu re 5 1 D ). Figure 5 to 5 4 below further show how populations of Aedes t aeniorhynchus will move back and forth starting from the coast to the eastern points in the county. The maps also illustrate how the population weighted centroids are a reasonable tool to describe movement patterns of Aedes t aeniorhynchus particularly when only mosquito management trap data is available. The pattern from the IDW maps tells a similar story to that of the population weighted mean center maps. Aedes t aeniorhynchus populations emerge along the coast then disperse inland. The seesaw nature of the map densities is indicative of population emerge nce and die off. Once a population emerges along the coast it will
37 disperse inland and die out. New populations will follow a similar cycle of emerging along the coast then dispersing east. The weighted centroids from April to August are shown in F igures 5 5, 5 6, and 5 7 for year 1992, 1993, and 1994, respectively. From each figure, a well defined east to west movement repetition can be observed. A clear example of this can be seen in the last four centroid locati ons in the 1992 map (Figure 5 5 ). Startin g with point 30 along the western coast I notice an eastern dispersal of the population going to point 31. There was a slight western pull back toward the coast with point 32 followed by a large eastern dispersal going to point 33 then 34 on the far easter n boarder of the county. The same patterns can also be identified in th e 1993 movement map (F igure 5 6 ). A good example of eastern dispersal is seen by starting at point 24 in the north west section of the map then move north east to point 25. The next poi nt, 26, moves south east towards the maps center. Point 27 is found to the east of 26 while 28 is on the far eastern edge of the map. This is a clear picture of population movement traveling east from the western coast. The poin ts depicted in 1994 (F igure 5 7 ) are also indicative of population movement. Looking at point 4 in the north west section the next count date, point 5, moves south east while point 6 travels south west back towards the coast. The next point, number 7, is found back east almost in the center of the map. Moving further east towards the edge of the map is where point 8 is found. The three points, 6 8, show an eastern dispersal that ranges from the western coast to the eastern section of Sarasota County. A possible explanation is the emer gence of new populations in salt marshes along the coast followed by a western dispersal. The movement pattern of the population weighted centroids and the pendulum like movement tells a story that follows
38 what I expected with this particular study species Aedes taeniorhynchus is a known coastal salt marsh breeder so it makes since to conclude that populations arise along the eastern coast of Sarasota County and then disperse inland to the west. This portrayal of mosquito movement shows the range and possi ble dangers associated with vector disease spread with Aedes taeniorhynchus The population weighted centroid maps for all three years show movement across the county. 1992 and 1993 show similar movement directions while 1994 seems to have more of a sharp depiction of west/east movement back and forth across the county. One hypothesis for this could be due to some environmental changes to the landscape such as the clearing of forested area or suburban development which could aid in mosquito movement. Perha ps looking into the effects of the El nino Southern Oscillation on mosquito populations could help explain the movement patterns seen. There was a moderately strong El nino event in 1994 so this could explain the slight differences seen in the mosquito mov ement directions during this time. Association between W ind Direction and Mosquito Moving D irection According to t he correlation analysis results in Figure 5 8 and 5 9 I found no correlation between the daily wind directions and between sequential movements of population weighted mosquito centroid locations. Figure 5 8 plots wind direction observations at the day before mosquito collection dates (DBC) against the moving direction of mosquito centroid at collection dates. The correlation analysis yie lded an which is extremely low considering a value of 1 = perfect corr elation. Figure 5 9 compares the mosquito moving direction to the wind direction at the date of the highest averaged wind speed (HAWS) between two mosquito collections. Correlation between the HAWS dates and centroid movement was quite low
39 with an I also separated the population weighted centroid data and only looked into the comparison of daily wind directional movement with only eastern movement of mosquito c entroid locations. This was an attempt to remove the western mosquito movement out of the scenario which is mostly due to new populations emerging along the coast pulling the centroids west. I defined an eastern as any movement along the positive x axis. A s before two sets of wind data were used. Using DBC movements with the eastern centroid movements yielded another very low value of (Figure 5 10 ). Similar results were found using the HAWS data set with eastern centroid movements. The comparison produced and (Figure 5 11 ). Assessing the results above, the data does not support a relationship of any kind between daily wind directional movement and the population weighted centroids. It could be that Aedes t aeniorhynchus populations in S arasota County are not dispersed by wind and other factors are at play in the role of dispersal. Other factors in mosquito ecology such as blood meal seeking or mating flights, regardless of wind directions, could be the intentional driver in Aedes t aenior hynchus populations in Sarasota County. The large geographic range of the mosquito collection data could be an impediment on the analysis of dispersal amongst Aedes t aeniorhynchus populations in Sarasota County. Daily wind movement and the temporal time frame of the mosquito collections could not accurately portray mosquito dispersal across the county as a whole. Future studies conducted on a smaller spatial and temporal scale could render different results.
40 Association between Daily Wind Speed and Mosquito Dispersal Distance Correlations between the wind speed and moving distance of population weighted centroids result in low values. When comparing movement between all sequen tial centroid locations from all three years to the days between the collection dates with the highest average wind speeds, the resulting value is (Figure 5 12 ). The correlation analysis between the extreme high wind speeds and the centroid movi ng distance results in a higher correlation but still not statis tically significant (Figure 5 13). Figure 5 14 examines the relationship between the extreme low wind speeds and the mosquito centroid moving distance, and the correlation analysis produces the lowest correlation yet with ( F igure 5 14 ). To remove any pull of the centroids back to the coast I only looked into distance associated between centroids on the positive x axis, that is to say eastern movements. When comparing all three years of eastern movement distances to wind speeds collected on the day between counts with the highest average the resulting correlation value is ( F igure 5 15 ). Lastly I explored the relationship between the distances for eastern movement and extreme high wind values. The correlation analysis result ed in the highest but still very low in terms of being significant ( F igure 5 16 ). Considering the low values between all correlations performed between wind speed values and distances between population weighted ce ntroid locations, I would fail to reject the null hypothesis that there is no relationship between the two. It is noteworthy to mention the inc rease in values between the extreme low wind to distance ( ) and the extreme high wind to eastern movement distance (
41 Even though both are not remotely close to a perfect correlation value of 1.0, the increase is notable. Given th e results between movement and wind speed values mentioned in this section, and wind direction values in the previous, it seems as though the dispersal of Ae des t aeniorhynchus populations in Sarasota County are goal oriented and not randomized flights dete rmined by wind. Association between Tide and Precipitation Levels and Mosquito Populations Figures 5 17, 5 18, 5 19 display the association between tide, precipitation, and Ae des t aeniorhynchus populations in year 1992, 1993, and 1994, respectively. For year 1992, the results of the correlation between tide and species abundance after the 7 day time lag are while the correlation value for precipitation are (Figure 5 17 ). The correlation amongst the tide levels for 1993 tell a simila r story with an but the precipitation value drops to an ( F igure 5 18 ). The results for 1994 yield similar values to that of the year before with tide at an and precipitation at ( F ig ure 5 19 ). Even though the values show fairly weak correlation, with the exception of precipitation to count during 1992, the charts show tide and precipitation are a good indicator if you concentrate on the peaks. A seven day time lag was placed on counts based on past literature review, but the life cycle of Ae des t aeniorhynchus is not exactly 7 days for each and every generation of the mosquito. Looking at the tide level spike in early June of 1992, you can see a spike in Ae des t aeniorhynchus numbers starting around the 10 th of t hat month. Very notable spikes in tide and precipitation levels occur at the end of June in 1992. On the 5 th of July you can see a very large increase in population numbers which seems to be a result of the high tide and precipitation level
42 during the end of June. 1993 shows a similar relationship between tide and counts. The largest spike in Ae des t aeniorhynchus population numbers on June 27 th occurs 5 7 occurred o n July 10 th and is not followed by a spike in Ae des t aeniorhynchus numbers. You can see that during this time tide levels surrounding July 10 th are very low. In 1994 you can see a large spike in tide levels peaking on May 26 th On June 8 th you can see a sp ike in population numbers. The next population spike can be seen on June 25 th and follows a period of high tide levels that start about 7 days before on the 18 th A spike in precipitation also occurred on the 16 th which could have helped elevate Ae des t aeniorhynchus numbers on the 25 th Based upon the date shown it would seem that high tide values are a better predictor for large Ae des t aeniorhynchus populations over precipitation. Females of this species are known to lay eggs in salt marshes along coa stal regions. When these regions are inundated with tidal waters the eggs are able to emerge and disperse. Limitations and Future Research Several limitations should be considered when int erpreting these results. T he temporal span of the mosquito collections was not even across all weeks throughout the years of the study. More systematic mosquito collections dates could facilitate additional space time analysis. Also, the IDW maps are isotropic which assumes all dir ections are the same. This is one example of a limitation while using this method. Future work could involve the use of Kriging as an interpolation technique in which the surrounding measured values are weighted to derive a predicted value for an unmeasure d location. Even though the use of a 7 day time lag is supported in the literature, a different numbers of days could produce higher values when comparing to precipitation and
43 tide. Another critique is that t he collection area for Sarasota County exclu des Myakka River State Park which is located in the center of the county. Information on mosquito populations is not known for this section of Sarasota County. Lastly, the geographical scale of analysis could be too large to properly analyze a relationship between wind and mosquito dispersal. Wind could only be a factor associated at distances less than the distance between centroid locations. Directions for future research could include a study performed from the analysis of Ae des t aeniorhynchus populations that have been systematically sampled at stationary trap locations for example, every day throughout the summer months in Florida. A study that has access to a full 360 degree sample area including traps located over the ocean would be ideal to compile a complete depiction of Ae des t aeniorhynchus population dispersal. I would also like to incorporate landscape type and human distribution into a future dispersal study. Urban environments and forested areas may affect the distance and direction at which mosquitoes disperse. City buildings and dense vegetation could affect prevailing wind directions and thus alter the path to which Ae des t aeniorhynchus populations will disperse. Sarasota County could benefit from the knowledge gained in this stud y. The population weighted centroids show a trend in movement from the coast towards inland so the implementation of more mosquito traps in this direction could be beneficial for collection data. Information on high population locations is very important t o effectively implement targeted control strategies against mosquito populations. I would also like to include more research in the future on the effects the El nino phenomenon has on mosquito populations and dispersal across Sarasota County.
44 Perhaps the 1994 event brought about events that created optimal conditions for an eastern dispersal from the coast that was not present in the previous two years.
45 A) June 10 B) June 14 C ) June 17 D ) June 21 Figure 5 1. IDW maps showing Aedes taeniorhynchus estimated population sizes in relation to the population weighted centroid location s The darker the color on the map, the larger the mosquito population Figure A, from June 10, 1992, shows population levels highest at the coat. Figure B, from June 14, 1992, shows t he population has moved inland. Figure C, from Jun 17, 1992, shows populations levels have been pulled back towards the coast. Figure D, from June 21, 1992, shows a movement inland once again The contour loge value is the logged count of mosquito populations while the count is the estimated number of mosquitoes.
46 A) June 28 B) July 1 C) July 5 D ) July 8 Figure 5 2. IDW maps showing Aedes taeniorhynchus estimated population sizes in relation to the population weighted centroid locations. The darker the color on the map, t he larger the mosquito population Figure A, June 21, 1992, shows the highest population levels inland from the coast. Figure B, July 1, 1992, shows density levels back along the coast. Fig ure C, July 5, 1992, shows the c ontinuation of highest populations along the coast. Figure D, July 8, 1992, shows an inland dispersal of numbers. The conto ur loge value is the logged count of mosquito populations while the count is the estimated number of mosquitoes.
47 A) July 12 B) July 15 C ) July 19 D ) July 22 Figure 5 3. IDW maps showing Aedes taeniorhynchus estimated population sizes in relation to the population weighted centroid locations. The darker the color on the map, the larger the mosquito population Figure A, July 12, 1992, shows s pecies numbers along the coast. Figure B July 15, 1992, shows population levels staying a long the coast. Figure C, July 19, 1992, shows a slight southern movement of t he centroid location and population levels. Figure D, July 22, 1992, shows a slight inland dispersal occurring. The contour loge value is the logged count of mosquito populations while the count is the estimated number of mosquitoes.
48 A) July 26 B) August 1 C) August 6 D ) August 9 Figure 5 4. IDW maps showing Aedes taeniorhynchus estimated population sizes in relation to the population weighted centroid locations. The darker the color on the map, the larger the mosquito population Figure A, July 26, 1992, shows populations numbers moving inland from the coast. Figure B, Au gust 1, 1992, Shows a population levels rising along the coast. Figure C, August 6, 1992, shows population levels dispersed inland from the previous count. Figure D, August 9, 1992, shows population locations rising back up along the coast again. The contour loge value is the logged count of mosquito populations while the count is the estimated number of mosquitoes.
49 Figure 5 5 Population weighted centroid map of Ae des taeniorhynchus for 1992 The arrows s how movement between centroid locations. The numbers indicate centroid locations over time.
50 Figure 5 6 Population weighted centroid map of Ae des taeniorhynchus for 1993 The arrows s how movement between cen troid locations The numbers indicate centroid locations over time.
51 Figure 5 7 Population weighted centroid map of Ae des taeniorhynchus for 1994 The arrows s how movement between centroid locations The numbers indicate centroid locations over time.
52 Figure 5 8 Wind direction in relation to mosquito moving direction Units are in degrees moving counter clockwise with 0 indicating due north. Figure 5 9 Wind di rection in relat ion to mosquito moving direction Units are in degrees moving counter clockwise with 0 indicating due north. R = 0.0225 0 50 100 150 200 250 300 350 400 0 100 200 300 400 Wind Direction in Degrees Mosquito Moving Direction in Degrees Day Before Count R = 0.007 0 50 100 150 200 250 300 350 400 0 100 200 300 400 Wind Drirection in Degrees Mosquito Moving Direction in Degrees Highest Average Wind Speed Between Trap Dates
53 Figure 5 10 Mosquito eastern movement to wind direction Units are in degrees moving counter clockwise with 0 indicating due north. Figure 5 11 Mosquito eastern movement to wind direction Units are in degrees moving counter clockwise with 0 indicating due north. R = 0.0199 0 50 100 150 200 250 300 350 400 0 100 200 300 400 Wind Direction in Degrees Mosquito Moving Direction in Degrees Eastern Movement Mosquito Direction BCD R = 0.0223 0 50 100 150 200 250 300 350 400 0 100 200 300 400 Wind Direction in Degrees Mosquito Moving Direction in Degrees Eastern Movement Highest Speed Day
54 Fi gure 5 12 1992 1994 Wind speed to distance between centroids Figure 5 13 Above average wind speed to moving distance between centroids R = 0.0103 0 2 4 6 8 10 12 14 0.00 5,000.00 10,000.00 15,000.00 20,000.00 Wind speed (miles /hour) Distance between Centroids (meters) 1992 94 Total Speed Distance R = 0.0399 0 2 4 6 8 10 12 14 0.00 5,000.00 10,000.00 15,000.00 20,000.00 Wind speed (miles/hour) Distance between Centroids (meters) 1992 94 Above Average Speed Distance
55 Figure 5 14 Below average wind speed to distance between centroids Figure 5 15 Eastern movement directions wind speed to distance R = 0.0004 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 0.00 5,000.00 10,000.00 15,000.00 20,000.00 Wind speed (miles/hour) Distance between Centroids (meters) 1992 94 Below Average Speed Distance R = 0.0204 0 2 4 6 8 10 12 14 0 5000 10000 15000 Wind speed (miles/hour) Distance between Centroids (meters) Eastern Movement Speed/Distance
56 Figure 5 16 Above average wind speed to eastern movement distance of studied mosquitoes R = 0.0839 0 2 4 6 8 10 12 14 0 2000 4000 6000 8000 10000 12000 14000 Wind Speed (Miles/hour) Distance between Centroids (meters) Above Average Speed/Distance Eastern Movement
57 Figure 5 17 Precipitation and tide to counts with 7 day lag shown on scatter plots, 1992
58 Figure 5 18 Precipitation and tide to counts with 7 day lag shown on scatter plots, 1993
59 Figure 5 19 Precipitation and tide to counts with 7 day lag shown on scatter plots, 1994
60 Table 5 1. Hypothesis testing for R 2 statistics under the null hypothesis that R 2 =0 Association R 2 F statistic P value Day before count wind direction Mosquito Movement direction (Figure 5 8) 0.0225 2.3 0.132 Highest average wind speed Mosquito movement (Figure 5 9) 0.007 0.705 0.403 Day before count wind E ast ern mosquito move ment (Figure 5 10) 0.0199 0.69 0.41 Highest average wind speed Mosquito move ment direction toward east (Figure 5 11) 0.0223 0.86 0.358 Wind S peed Mosquito moving distance (Figure 5 12) 0.0103 1.527 0.219 Above Average Wind speed M osquito moving distance (Figure 5 13) 0.0399 2.11 0.154 Below Average speed Mosquito distance (Figure 5 14) 0.0004 0.388 0.536 Eastern movement speed Mosquito distance (Figure 5 15) 0.0204 0.94 0.337 Above Average speed Mosquito distance eastern movement (Figure 5 16) 0.0839 2.4 0.137 Tide Mosquito count 1992 (Figure 5 17) 0.18 2.12 0.124 Precip itation Mosquito count 1992 (Figure 5 17) 0.26 3.2 0.097 Tide Mosquito count 1993 (Figure 5 18) 0.12 2.11 0.156 Precip itation Mosquito count 1993 (Figure 5 18) 0.0056 0.388 0.403 Tide Mosquito count 199 4 (Figure 5 19) 0.1175 2.34 0.105 Precip itation Mosquito count 1994 (Figure 5 19) 0.0372 2.12 0.154
61 CHAPTER 6 CONCLUSION S The primary goal of this thesis was to explore the spatio temporal dynamics of the Ae des t aeniorhynchus mosquito in Sarasota County, Florida. Population weighted centroids from Sarasota mosquito management trap sites were calculated and used to show locations of density and movement across the study region. Distances and movements were calculated between se quential centroid locations and then compared with daily wind directions and speed. Daily tide and precipitation levels were compared to trapped Ae des t aeniorhynchus population numbers to analyze correlations between the two. There are three major findings from this research: 1) Ae des taeniorhynchus populations emerge along the coast then disperse inland as seen from the population weighted centroid maps. The population weighted centroid maps are found to be a reasonable tool to describe the movement patte rns, particularly when only mosquito management trap data is available. 2) There is no statistical correlation found between Ae des taeniorhynchus dispersal and daily wind speed, as well as direction s in Sarasota County, Florida Correlations between wind data and the population weighted centroids using Values were extremely low. Previous studies also show similar results to my findings. In a mark release recapture study carried out in New Hampshire on the tree hole mosquito Ochlerotatus triseriatus recaptues in the study were not related to the prevailing wind direction [ 40 ] Another mark release recapture study from Australia involving Aedes aegypti which is in the same genus as my study species, indicates a statistically significant trend for upwind flight [ 41 ]
62 3) There are stronger relationship s between Ae des t aeniorhynchus population size and tide levels, as well as precipitation levels This result is supported by previous findings which show population spikes of Ae des t aeniorhynchus after periods of high tide or precipitation levels [8 ], [ 10]. Another reason is that f emales often lay eggs in the ground and hatch 2 3 days after being flooded by rain or tide [ 11 ] Arboviruses are of great concern to the state of Florida. Disease such as WNV, SLEV, and EEE are found throughout the state. With international travel on the rise and possible introduction of other arboviruses, Florida residents are facing an elevated risk of imported mosquito vectored illness. Knowledge of the temporal and spatial distribution of populations of potentially impo rtant disease carrying mosquito vectors is important for categorizing areas of varying risk for disease transmission. Since Aedes taeniorhynchu s is a coastal species with the capacity to transmit indigenous and exotic arboviruses, and there is a potential for an introduction of exotic diseases into the United States through shipping ports, enhanced surveillance and control measures need to be established. My thesis is one such example of vector surveillance and show the capacity for studying aerial movement of disease carrying vectors.
63 LIST OF REFERENCES 1. Godsey MS, Jr., Blackmore MS, Panella NA, Burkhalter K, Gottfried K, Halsey LA, Rutledge R, Langevin SA, Gates R, Lamonte KM et al : West Nile virus epizootiology in the southeastern United States, 2001 Vector Borne Zoonotic Dis 2005, 5 (1):82 89. 2. [ http://www.cdc.gov/ncidod/dvbid/westnile/surv&control_archive.htm ] 3. St. Louis Encephalitis [ http://www.doh.state.fl.us/Environment/medicine/arboviral/StLouis.html ] 4. Eastern Equine Encephalitis [ http://www.doh.state.fl.us/Environment/medicine/arboviral/EasternEquine.html ] 5. Seth C. Britch KJL, Assaf Anyamba, Compton J. Tucker, Edwin W. Pak, Francis A. Maloney, Kristin Cobb, Erin Stanwix, Jeri Humphries, Alexandra Spring, Benedict Pagac, Melissa Miller: Satellite Vegetation Index Data as a Tool to Forecast Population Dynamics of Medically Important Mosquitoes at Military Installations in the Continental United States Military Medicine 2008, 173 (7):679 683. 6. Ailes MC: Failure to predict abundance of saltmarsh mosquitoes Aedes sollicitans and A. taeniorhynchus (Diptera: Culicidae) by using variables of tide and weather J Med Entomol 1998, 35 (3):200 204. 7. Turell MJ, Dohm DJ, Mores CN, Terrac ina L, Wallette DL, Jr., Hribar LJ, Pecor JE, Blow JA: Potential for North American mosquitoes to transmit Rift Valley fever virus J Am Mosq Control Assoc 2008, 24 (4):502 507. 8. SETH C. BRITCH KJL, ASSAF ANYAMBA, COMPTON J. TUCKER, EDWIN W. PAK: LONG TE RM SURVEILLANCE DATA AND PATTERNS OF INVASION BY AEDES ALBOPICTUS IN FLORIDA Journal of the American Mosquito Control Association 2008, 24 (1):115 120. 9. APPERSON C: The black salt marsh mosquito, Aedes taeniorhynchus Wing Beats 1991, 4 (4):9. 10. Erik Tetens Nielsen ATN: Field Observations on the Habits of Aedes Taeniorhynchus Ecology 1953, 34 (1):141 156. 11. Scott A. Ritchie CLM: Simulated populations of the black salt marsh mosquito ( Aedes taeniorhynchus) in a Florida mangrove forest Ecological Modelling 1995, 77 :123 141.
64 12. Williams CB: Studies on the effect of weather conditions on the activity and abundance of insect populations Phil Trans R Soc Lond B 1961, 244 (713):331 378. 13. Service M: Field Sampling Methods In: Mosquito Eco logy vol. 2nd edn. London: Chapman & Hall; 1993. 14. Nayar J: Measuring Adult Dispersal In: Mosquito Ecology Springer 2008: 1377 1424. 15. Hamlyn Harris R: Some ecological factors involved in the dispersal of mosquitoes in Queensland Bull Entomol Res 1933, 24 :229 232. 16. D. MacCreary LS: Mosquito Migration Across Delaware Bay Proc New Jers Mosq Exterm Assoc 1937, 24 :188 197. 17. Curry D: A documented record of a long flight of Aedes sollicitans Proc New Jers Mosq Exterm Assoc 1939, 26 :36 39. 18. Kohn GC (ed.): Encyclopedia of Plague and Pestilence from Ancient Times to the Present Third edition edn. New York: Facts on File 1995. 19. Tatem AJ, Hay SI, Rogers DJ: Global traffic and disease vector dispersal Proc Natl Acad Sci U S A 2006, 103 (16):6242 6247. 20. Clark D: An analysis of dispersal and movement in Phaulacridium vittatum Australian Journal of Zoology 1962, 10 (3):382 399. 21. Sithiprasasna R, Linthicum KJ, Liu GJ, Jones JW, Singhasivanon P: Use of GIS based spatial modeling approa ch to characterize the spatial patterns of malaria mosquito vector breeding habitats in northwestern Thailand Southeast Asian J Trop Med Public Health 2003, 34 (3):517 528. 22. Dale PE, Ritchie SA, Territo BM, Morris CD, Muhar A, Kay BH: An overview of re mote sensing and GIS for surveillance of mosquito vector habitats and risk assessment J Vector Ecol 1998, 23 (1):54 61. 23. Shone SM, Curriero FC, Lesser CR, Glass GE: Characterizing population dynamics of Aedes sollicitans (Diptera: Culicidae) using mete orological data J Med Entomol 2006, 43 (2):393 402. 24. Takeda T, Whitehouse CA, Brewer M, Gettman AD, Mather TN: Arbovirus surveillance in Rhode Island: assessing potential ecologic and climatic correlates J Am Mosq Control Assoc 2003, 19 (3):179 189.
65 2 5. Woodruff RE, Guest CS, Garner MG, Becker N, Lindesay J, Carvan T, Ebi K: Predicting Ross River virus epidemics from regional weather data Epidemiology 2002, 13 (4):384 393. 26. Haeger JS: Behavior preceding migration in the salt marsh mosquito, Aedes t aeniorhynchus (Wiedemann). Mosquito News 1960, 20 (2):136 147. 27. Sankari T, Hoti SL, Singh TB, Shanmugavel J: Outbreak of dengue virus serotype 2 (DENV 2) of Cambodian origin in Manipur, India Association with meteorological factors Indian J Med Res 2 012, 136 (4):649 655. 28. Anyamba A, Linthicum KJ, Small JL, Collins KM, Tucker CJ, Pak EW, Britch SC, Eastman JR, Pinzon JE, Russell KL: Climate teleconnections and recent patterns of human and animal disease outbreaks PLoS Negl Trop Dis 2012, 6 (1):e1465. 29. Deichmeister JM, Telang A: Abundance of West Nile virus mosquito vectors in relation to climate and landscape variables J Vector Ecol 2011, 36 (1):75 85. 30. Landesman WJ, Allan BF, Langerhans RB, Knight TM, Chase JM: Inter annual associations between precipitation and human incidence of West Nile virus in the United States Vector Borne Zoonotic Dis 2007, 7 (3):337 343. 31. Rey JR, O'Meara GF, O'Connell SM, Cutwa Francis MM: Factors affecting mosquito production from stormwater dra ins and catch basins in two Florida cities J Vector Ecol 2006, 31 (2):334 343. 32. Douglas J. Pool SCS, Ariel E. Lugo: Structure of Mangrove Forests in Florida, Puerto Rico, Mexico, and Costa Rica Biotropica 1977, 9 (3):195 212. 33. Chapman JW, Nesbit RL Burgin LE, Reynolds DR, Smith AD, Middleton DR, Hill JK: Flight orientation behaviors promote optimal migration trajectories in high flying insects Science 2010, 327 (5966):682 685. 34. D.E. Pedgley DRR, G.M. Tatchell: Insect Migration: Tracking Resourc es through Space and Time : Cambridge University Press; 1995. 35. ason W. Chapman DRR, Alan D. Smith, Joe R. Riley, David E. Pedgley, Ian P. Woiwod: High altitude migration of the diamondback moth Plutella xylostella to the U.K.: a study using radar, aeria l netting, and ground trapping Ecological Entomology 2002, 27 (6):641 650. 36. Walker TJ: Migration and re migration of butterflies tlirough north peninsular Florida: quantification with malaise traps. J Lepid Soc 1978, 32 :178 190.
66 37. Glick PA: The Distribution of Insects, Spiders, and Mites in the air US Dep Agric Tech Bull 1939, 673 :149. 38. V. A. Drake KFH, J. L. Readshaw, D. G. Reid Insect migration across Bass Strait, Australia during spring; a radar study. Bull Entomol Res 1981, 71 :449 466 39. Ingrid H. Williams DF, Hassan Barari, Alastair McCartney: Migration to and dispersal from oilseed rape by the pollen beetle, Meligethes aeneus, in relation to wind direction Agricultural and Forest Entomology 2007, 9 (4):279 286. 40. Ellis AM: Link ing movement and oviposition behaviour to spatial population distribution in the tree hole mosquito Ochlerotatus triseriatus J Anim Ecol 2008, 77 (1):156 166. 41. Russell RC, Webb CE, Williams CR, Ritchie SA: Mark release recapture study to measure disper sal of the mosquito Aedes aegypti in Cairns, Queensland, Australia Med Vet Entomol 2005, 19 (4):451 457. 42. Florida Mosquito Control [ http://mosquito.ifas.ufl.edu/Florida_Mosquito_ Control.htm ] 43. Mosqutio Management [ http://www.sarasotahealth.org/services/environmental mosquito.htm ] 44. REINERT WC: THE NEW JERSEY LIGHT TRAP: AN OLD STANDARD FOR MOST MOSQUITO CONTROL PROGRAMS In: Proceedings of the Seventy Sixth Annual Meeting of the New Jersey Mosquito Control Association, Inc 1989: 17 25. 45. About Sarasota County [https:// www.scgov.net/AboutSarasota/Pages/default.aspx ] 46. National Data Buoy Center [ http://www.ndbc.noaa.gov/maps/florida_hist. shtml ] 47. Neela Thapar DW, Jay Lee: The Changing Geography of Population Centroids in the United States between 1970 and 1990 The Geographical Bulletin 1999, 41 (1):45 57. 48. Paul J. Boyle RF: Improving Distance Estimates between Areal Units in Migrati on Models Geographical Analysis 1997, 29 (2):93 107. 49. Paul J. Boyle RF: Migration Trends for the West Midlands: Suburbanisation,Counterurbanisation or Rural Depopulation? Migration Processes and Patterns 1992, 2 :44 61.
67 50. Melinda S. Meade RJE: Medical Geography Second Edition edn: The Guilford Press; 2005. 51. Tachiiri K, Klinkenberg B, Mak S, Kazmi J: Predicting outbreaks: a spatial risk assessment of West Nile virus in British Columbia Int J Health Geogr 2006, 5 :21. 52. Thomas R. Allen BS: Spatial interpolation and image integrative geostatistical prediction of mosquito vectors for arboviral surveillance Geocarto International 2008, 23 (4):311 325. 53. Lance A. Waller CAG: Applied Spatial Statistics for Public Health Data Wiley; 2004. 54. The Geography of Transport Systems [ http://people.hofstra.edu/geotrans/eng/ch1en/conc1en/greatcircle.html ] 55. Aviation Formulary V1.46 [ http://williams.best.vwh.net/avform.htm ] 56. Shaman J, Stieglitz M, Stark C, Le Blancq S, Cane M: Using a dynamic hydrology model to predict mosquito abundances in flood and swamp water Emerg Infect Dis 2002, 8 (1):6 13. 57. Nayar JK: Bionomics and Physiology of Aedes taeniorhynchus and Aedes sollicitans, the salt marsh mosquitoes of Florida. Bulletin, Agricultural Experiment Stations, University of Florida 1985, 852
68 BIOGRAPHICAL SKETCH the geography department he worked on mosquito research for the United States Departme nt of Agriculture in Gainesville, Florida. After graduation he plans to continue is travels around the globe and enter a PhD program studying vector disease.