INFORMATION TO USERS This material was produced from a microfilm copy of the original document. While the most advanced technological means to photograph and reproduce this document have been used, the quality is heavily dependent upon the quality of the original submitted. The following explanation of techniques is provided to help you understand markings or patterns which may appear on this reproduction. 1.The sign or "target" for pages apparently lacking from the document photographed is "Missing Page(s)". If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting thru an image and duplicating adjacent pages to insure you complete continuity. 2. When an image on the film is obliterated with a large round black mark, it is an indication that the photographer suspected that the copy may have moved during exposure and thus cause a blurred image. You will find a good image of the page in the adjacent frame. 3. When a map, drawing or chart, etc., was part of the material being photographed the photographer followed a definite method in "sectioning" the material. It is customary to begin photoing at the upper left hand corner of a large sheet and to continue photoing from left to right in equal sections with a small overlap. If necessary, sectioning is continued again — beginning below the first row and continuing on until complete. 4. The majority of users indicate that the textual content is of greatest value, however, a somewhat higher quality reproduction could be made from "photographs" if essential to the understanding of the dissertation. Silver prints of "photographs" may be ordered at additional charge by writing the Order Department, giving the catalog number, title, author and specific pages you wish reproduced. 5. PLEASE NOTE: Some pages may have indistinct print. Filmed as received. Xerox University Microfilms 300 North Zeeb Road Ann Arbor, Michigan 48106 I I I I 75-20,899 WAGNER, Vaughn Edwin, Jr., 1942MOSQUITO BITING ACTIVITY IN MICHIGAN STATE PARKS AND DEVELOPMENT OF A SYSTEMS SCIENCE MODEL FOR STUDY OF WOODLAND AEDES. Michigan State University, Fh.D0, 1975 Entomology Xerox University Microfilms, Ann Arbor, M ichigan 48106 MOSQUITO BITING ACTIVITY IN MICHIGAN STATE PARKS AND DEVELOPMENT OF A SYSTEMS SCIENCE MODEL FOR STUDY OF WOODLAND AEDES By Vaughn E. Wagner A DISSERTATION Submitted to Michigan State U niversity in p a rtia l fu lfillm e n t of the requirements fo r the degree o f DOCTOR OF PHILOSOPHY Department o f Entomology 1975 ABSTRACT MOSQUITO BITING ACTIVITY IN MICHIGAN STATE PARKS AND DEVELOPMENT OF A SYSTEMS SCIENCE MODEL FOR STUDY OF WOODLAND AEDES By Vaughn E. Wagner Mosquito b itin g a c tiv ity was studied in the Michigan park system from 1971 to 1973. Results indicated th at woodland Aedes mosquitoes were the problem species in the m ajority of state parks. Adult collections made a t North Higgins Lake and Yankee Springs parks in 1972 and 1973 revealed that mosquitoes of the JL communistrichurus and _A. s tim u la n s -fitc h ii species complexes, resp ectively, were the major pests in these two parks. The former complex was a problem a t N. Higgins Lake Park with highest b itin g a c tiv ity occurring la te May and e a rly June. The la t t e r complex was a problem at Yankee Springs Park during most of the summer with b itin g a c tiv ity extending u n til mid-August. Highest a c tiv ity fo r the two-year study occurred during June and July. The sampling t r ia ls at each park were conducted a t three areas designated fo r public use and at peak b itin g levels recreational a c tiv itie s were cancelled by park personnel. Field investigations into sp ecific areas of A_. stimulans- f i tc h ii population dynamics was also conducted. A representative breeding s ite was surveyed during 1972 and 1973 and data on in star Vaughn E. Wagner composition and developmental rates were collected. on adult female longevity was also obtained. Information Immature and adult m o rta lity rates were estimated fo r a population cohort, larval densities were calculated, and female egg input response was redefined. Research d e a lt, in p a rt, with the construction o f a system model which gave a description of the dynamic behavior o f this insect group. This approach was e s s e n tia lly a sequential structuring process involving the follow ing: a d e fin itio n of the system and its components; determination o f the behavioral features and th e ir causal orientations fo r each component o f the system; the constraints by which the components are governed; and free body modeling of each component. Since the model w ill be used as an aid in developing la rv a l control s tra te g ie s , emphasis was placed on the interaction between the aquatic environment and the immature l i f e stages. As a re s u lt behavioral features involving these interactions were chosen fo r inclusion in the model. Features of the aquatic environment were viewed as stim uli to the biological components of in te re s t while the response variables were aspects o f the Aedes l i f e cycle which would r e fle c t the e ffe cts o f control measures. A computer simulation implementing th is model was constructed and refined u t iliz in g results from population studies of A: stimulansf i tc h ii mosquitoes. Information and data u tiliz e d in the model construction were obtained from surveys conducted in Michigan state parks and an extensive lite r a tu r e review. Dynamics of the mosquito's l i f e cycle and o f the woodland pool ecosystem were modeled and the Vaughn E. Wagner resu ltan t equations are described. Assumptions used in the modeling process are b r ie fly discussed in order to show the reasoning process underlying the use o f these equations in the sim ulation. A description of computer program structure and data f ile s u tiliz e d by the program is given. Research on and refinement of the physical components consisted of measurements during 1973 on aquatic parameters chosen as key factors in immature mosquito dynamics and meteorological features oriented as stim uli to these parameters. The resu ltan t data were s t a t is t ic a lly analyzed to obtain the best av ailab le model fo r maximum water temperature, dissolved oxygen concentration, and water depth. ACKNOWLEDGMENTS I wish to express my gratitude to Dr. H.D. Newson fo r his support and guidance throughout the course of th is study. His ideas and c r it ic a l review o f the research methods are sincerely appreciated. Thanks to Dr. E.D. Goodman fo r providing the time and assistance in formulating the concepts necessary fo r a systems science approach to th is entomological problem. I express my appreciation to the members of my graduate guidance committee, Drs. H.D. Newson, E.D. Goodman, K.W. Cummins, J.E. Bath, and D.L. Haynes. The opportunity to associate w ith these distinguished scie n tis ts has been an enlightening experience. I wish to thank those individuals of the Michigan State Department o f Natural Resources fo r th e ir assistance in the f ie ld surveys, especially to park personnel a t North Higgins Lake and Yankee Springs. I extend thanks also to Mr. G.A. T u lly , a very able and knowledgeable computer programmer, who was instrumental in constructing the overall program fo r the simulation model. Most important, profound appreciation is extended to my dear w ife , S hirley Marie, whose devotion, patience, and support made the attainment o f th is doctoral degree possible. ii TABLE OF CONTENTS Page LIST OF TABLES...................................................................................................... v LIST OF FIG U R E S .................................................................................................. vi ...................................................................................... 1 GENERAL INTRODUCTION PART I SURVEY OF MOSQUITO BITING ACTIVITY IN MICHIGAN STATE PARKS AND A STUDY OF AEDES STIMULANS-FITCHII POPULATION DYNAMICS INTRODUCTION .......................................................................................................... 6 SURVEY METHODS ...................................................................................................... 7 Adult Mosquito Studies . Immature Mosquito Studies .................................................................. .................................................................. RESULTS......................................................................... Adult Mosquito Collections .................................................................. Comparison o f B iting A c tiv ity and E ffe c t on Recreation . . . Dynamics o f a Woodland Aedes Species Complex ............................. Immature and Adult M o rta lity .............................................................. 7 8 9 9 12 14 16 PART I I A SYSTEMS SCIENCE APPROACH TO THE STUDY OF BITING INSECT POPULATIONS IN MICHIGAN STATE PARKS INTRODUCTION ........................................................................................................... 35 DEFINITION OF THE AEDES ECOSYSTEM .............................................................. 37 O rientation o f Behavioral Features into a Stimulus and Response S e t .................................................................................. Construction of Free Body Models ...................................................... 39 41 iii Page PART I I I CONSTRUCTION AND REFINEMENT OF A COMPUTER SIMULATION MODEL FOR POPULATION STUDIES OF WOODLAND POOL AEDES MOSQUITOES INTRODUCTION .......................................................................................................... 55 MODELING THE BIOLOGICAL COMPONENTS .............................................................. 56 Dynamics o f Dynamics o f Dynamics of the Immature Stages .................................................... the Egg S t a g e ................................................................ the Adult S t a g e ........................................................... MODELING THE PHYSICAL Dynamics of PROGRAM STRUCTURE 58 61 63 COMPONENTS ................................................................ 65 ........................................... 66 .............................................................................................. 71 the Physical Components REFINEMENT OF THE MODEL'S BIOLOGICAL COMPONENTS ................................. 73 Immature and Adult M o rta litie s Used in Model ............................. Female Gonotrophic Cycle ...................................................................... 73 75 REFINEMENT OF PHYSICAL COMPONENTS ................. 78 Water T e m p e ra tu re .................................................................................. Dissolved Oxygen Concentration .......................................................... Water D e p t h .............................................................................................. 78 79 80 SUMMARY AND CONCLUSION ...................................................................................... 91 LITERATURE CITED .................................................................................................. 96 I LIST OF TABLES Table Page PART I 1. 2. 3. 4. 5. 6. 7. Percent composition o f the major mosquito species in la n d in g /b itin g collections made in ten state parks, 1 9 7 1 .......................................................................................... 19 B iting a c tiv ity of Aedes communis- trichurus complex a t North Higgins Lake state park, 1972-1973 ......................... 20 B iting a c tiv ity of Aedes stimulans- f i t c h i i complex at Yankee Springs recreational area, 1972-1973 21 Immature Aedes stimulans- f i t c h i i in a woodland pool a t Yankee Springs recreational area, 1972 ............................ 22 Immature Aedes stimulans- f i t c h i i in a woodland pool a t Yankee Springs recreational area, 1973 ............................ 23 L ife tables fo r immature Aedes stimulans- f i t c h i i in a woodland pool ecosystem a t Yankee Springs recreational area, 1972-1973 ..................................................... 24 L ife tables fo r adult Aedes stimulans- f i t c h i i at Yankee Springs recreational area, 1972-1973 ......................... 25 PART I I 1. 2. Behavioral features associated with components of Aedes ecosystem with emphasis on immature-aquatic interactions ...................................................................................... 46 State and response equation forms generated by the free body models o f each component in the Aedes s y s t e m ...................................................................................... 47 v Page Table PART I I I 1. 2. Age structure o f a population of un ivoltine Aedes m o s q u ito e s .......................................................................................... 83 C orrelation m atrix fo r max-min a ir and water temperatures and dissolved oxygen concentration in a woodland pool ecosystem a t Yankee Springs recreational area, 1973 .................................................................. 84 vi LIST OF FIGURES Figure Page PART I 1. 2. 3. 4. Park and recreational areas in Michigan where mosquito surveys were conducted during 1971 ............................ 27 Percent species composition o f the Aedes snowpool mosquitoes id e n tifie d from collections made in nine Michigan state park and recreational areas, 1971 29 Average mosquito b itin g a c tiv ity and park attendance fo r 1972 and 1973 a t North Higgins Lake state park . . . . 31 Average mosquito b itin g a c tiv ity and park attendance fo r 1972 and 1973 at Yankee Springs recreational area . . . 33 PART I I 1. 2. 3. A diagram o f woodland Aedes population components and th e ir respective environments . . . ......................................... 49 System graph o f Aedes mosquito ecosystem indicating stimulus Sbj and response Rb-j, o rie n tatio n of the behavioral features .......................................................................... 51 Construction of free body models associated with the Aedes com ponents.................................................................................. 53 PART I I I 1. 2. 3. Flowchart o f Aedes population dynamics as expressed by the computer simulation model ................................................. 86 Flowchart of the program structure of the Aedes computer simulation model ................................................................. 88 Survivorship curves fo r Aedes s tim u la n s -fitc h ii adults at Yankee Springs recreational area, 1972-1973 .................................................................................................. 90 vi 1 ! GENERAL INTRODUCTION A research program was proposed by representatives of the Michigan Department o f Natural Resources to study mosquito and other biting insect problems o f state parks and recreational areas. The proposal o f th is study was due to concern that existin g mosquito control e ffo rts in park areas merited a complete reevaluation. Unsuccessful attempts in the past to regulate these b itin g insects made i t imperative th at studies on species composition, r e la tiv e abundance, and seasonal d is trib u tio n be undertaken. As temporary woodland pools in the park areas are u tiliz e d as breeding sites by certain species of Aedes mosquitoes, a research project concerning interaction of the immature l i f e stages with these aquatic ecosystems was also proposed. Previous research (H orsfall e t a l . , 1958; H o rs fa ll, 1963; Horsfall and T rp is, 1967) has shown th a t inundation of the breeding s ite s , a threshold water temperature, and reduced oxygen tension are necessary stim uli fo r fin a l egg hatch. Hatching in the f ie ld occurred when the water temperature rose above 5° C. The dynamics o f decreasing oxygen content in a temporary woodland pool is not well understood (Ruttner, 1963). In the la rv al stage the influence o f water temper­ ature on the rate o f development has been researched by Anderson and Horsfall (1969) and Trpis and H orsfall (1969). 1 Results indicated th at 2 the range of temperature th at promotes normal growth fo r two species of woodland Aedes mosquitoes was 5° to 21° C. Length o f larval development as well as in s ta r duration depended on water temperature. Any temperature in excess o f 21° C had an adverse e ffe c t on the maturing mosquito e ith e r in increased m o rta lity or eventual supression of male c h a ra c te ris tic s . McDaniel (1958) observed th at development from hatching to emergence o f A. stimulans required 13 to 60 days, depending on temperature. Biological control o f JL stimulans in woodland pools was researched by James (1961) who observed th at adult dytiscids and immature Limnephilis sp. were predators o f early in s ta r larvae. Mermithid nematode parasitism o f the same mosquito species complex was studied by Hagan (1966) and Dayton (1971). Adults of woodland Aedes mosquitoes are found in highest density adjacent to pool areas and w ill disperse 200 to 300 yards (McDaniel, 1958). D e fo lia rt e t a l. (1967) has reported th at b itin g a c tiv ity fo r thejA. stimulans complex in Wisconsin began in mid-May, peaked during June, and existed through July and August. Hagan (1966) reported s im ila r b itin g a c tiv ity fo r th is species complex in Michigan. Woodland Aedes females i n it ia t e b itin g a c tiv ity 10 to 15 days a fte r emergence with successive portions of the population laying eggs throughout the season u n til August (Detinova, 1968). Eggs are deposited on moist layers of plant d e tritu s along the flood lin e o f woodland pool micro­ basins (H o rs fa ll, 1963). The woodland Aedes group as represented by the_A. stimulans and_A. communis species complexes are univoltine with egg diapause shown to be obligatory (H orsfall and Fowler, 1961). Adult 3 d istrib u tio n fo r the Aedes woodland group is throughout Michigan (Pedersen, 1947) with the A. stimulans complex most abundant in the southern lower peninsula and A_. communis complex more prevalent in northern Michigan (Irw in , 1942). A recent study by Gorton (1973) showed that woodland Aedes mosquitoes were collected in substantial numbers a t a southern Michigan subdivision. Due to the sheer complexity of the relatio n sh ip between the aquatic environment and immature l i f e stages o f woodland Aedes mosquitoes, a systems science approach was undertaken. Systems ecology has shown that ecosystems can be modeled u sefu lly and the value of the modeling process ju s t if ie s the e f fo r t . The re s u lt has been a coordinating e ffe c t on organizing research projects as well as c la rify in g components o f an ecosystem, id e n tify in g data require­ ments and determining research p r io r itie s . Given a model developed according to th is approach and u t iliz in g requirements of these mosquitoes, a computer simulation was one of the methods availab le fo r examining behavior and e ffe c ts of external s tim u li. The weak­ ness of many computer simulation studies is th a t they provide l i t t l e more than a tabulation of p a rtic u la r or overall behavior of the system. The use o f systems science in imparting an intern al structure to the simulation model can re s u lt in the elim ination o f such defects. As a re s u lt, a framework could be provided upon which to base decisions concerning steps to be taken in a lte rin g the system to obtain adequate pest management o f th is mosquito. The accuracy of these predictions would depend on how well the simulated model represented biological 4 r e a lity . Continual comparison with the natural ecosystem would be required to adequately valid a te the simulation model. In view o f these considerations, a research project was conducted in the Michigan park system during 1971-1973 with the following objectives: (1) conduct mosquito surveys in the Michigan state park system; (2) study the relation ship between the aquatic environment and immature l i f e stages of woodland Aedes mosquitoes; (3) construct a simulation model o f the woodland Aedes ecosystem u tiliz in g a systems science approach. PART I SURVEY OF MOSQUITO BITING ACTIVITY IN MICHIGAN STATE PARKS AND A STUDY OF AEDES STIMULANS- FITCHII POPULATION DYNAMICS1 INTRODUCTION Adult mosquito surveys were conducted in selected Michigan state parks during 1971-1973. The primary objectives were to deter­ mine the human-biting species present a t public use areas and to study the duration and in te n s ity o f mosquito b itin g a c tiv it y . The need fo r studies on species composition, re la tiv e abundance, and seasonal d is trib u tio n was emphasized due to increased awareness of potential arthropod disease a c tiv ity as well as widespread disruption of human recreational a c tiv itie s . Studies were also conducted during 1972 and 1973 on populations of the Aedes stimulans- f i t c h i i species complex. An important consider­ ation in understanding population dynamics of th is mosquito was comprehension of immature stage--aquatic breeding s ite in teractio n s. I f these interactions were b e tte r understood, then insights in to the regulation of population densities could be elucidated. Accordingly, biological analysis would provide a framework upon which to base decisions concerning strategies to be taken in a lte rin g the Aedes ecosystem to obtain adequate management o f th is pest. i 6 i et IL... SURVEY METHODS Adult Mosquito Studies In 1971 eleven park areas (Fig. 1) in Michigan were chosen fo r mosquito surveys that included adult landing and b itin g collectio n s. Park natu ralists conducted the surveys in public use areas w ithin each park. The collections were sent to the medical entomology laboratory at Michigan State University fo r id e n tific a tio n . Based upon th is survey North Higgins Lake and Yankee Springs, recreation areas located in north central and southern Michigan, resp ectively, were selected fo r fu rth e r studies. B iting collections were continued a t these two parks in 1972 and 1973. North Higgins Lake Park is located on the northern end o f Higgins Lake in a hardwood and pine forested area and is 428.94 acres in size. The Yankee Springs rec­ reational area is located on Gun Lake and comprises 4,971.57 acres in a maple and oak hardwood fo rest region. The b itin g count collections were standardized as follows: 1. Collections were started with the emergence o f the Aedes snowpool group and conducted every seventh day u n til the mean b itin g count was <10 per 10 minute sampling period fo r two consecutive sampling t r ia ls . 2. Collections were made a t three public use areas: nature t r a i l , campground, and outdoor education center. 7 8 3. Collections were made during the morning hours between 8:00-10:00 a.m. 4. Mosquitoes collected were e ith e r probing or ac tu a lly feeding on the c o lle c to r. Immature Mosquito Studies F ield research was also conducted during 1972 and 1973 on immature mosquito populations a t Yankee Springs recreational area in southwestern Michigan. Investigations into specific areas of population dynamics was indicated by a lack o f data and/or informa­ tion concerning th is species complex's ecosystem. As a prelim inary survey of four woodland pools in the park indicated uniform la rv a l densities among pools, a representative breeding s ite was chosen fo r in-depth surveys (Wagner).2 I t consisted o f an oval microbasin area subject to seasonal, temporary flooding with maximum flood occurring in early spring. This flooding produced an estimated surface area of 2,364 sq. f t . and provided su itab le h ab ita t fo r the hatch and development o f the immature stages. Larvae and pupae were collected during and a fte r periods of maximum flood along four transect lines at rig h t angles to the long axis of the pool. During 1972, samples were taken a t 7 day in te rv als beginning A p ril 14 and ending May 10. A to ta l o f 24 water samples representing 2.73 sq. f t . o f surface area was taken a t each t r i a l . In 1973 the procedure was repeated except that 60 samples were taken a t each t r i a l (representing 8.18 sq. f t . of surface area) and the sampling period extended from March 27 to A pril 24 due to e a r lie r egg hatch. RESULTS Adult Mosquito Collections A to ta l o f 848 adu lt mosquitoes submitted from the 10 state parks were id e n tifie d (no adu lt specimens were submitted from North Higgins Lake) in 1971 to determine the species composition and f r e ­ quency of occurrence throughout the park system. Table 1 lis t s the species of mosquitoes id e n tifie d from landing and b itin g collections made in the 10 parks during 1971. The Aedes snowpool group was the most abundant species group collected in seven parks (range 40.0% in park 8 to 90.2% in park 2) while C o q u ille ttid ia perturbans constituted 76.5 and 92.3 percent in parks 9 and 11 and Anopheles quadrimaculatus comprised 46.4 percent o f the sample from park 5. Aedes vexans, Aedes tris e ria tu s and Culex pjpiens made up the re ­ maining portion of the sample population. Frequency of occurrence and the percentage o f the to ta l number o f collection s in which a certain species or species group occurs is also shown in Table 1. The Aedes snowpool group and JC. perturbans had the highest frequency of occurrence, being present in 90 percent of the park samples. During the same c o lle c tio n periods A. vexans was present in 50 percent o f the samples w hile k. quadrimaculatus and A. tris e ria tu s showed a frequency of occurrence o f 40 percent _C. pipiens had the lowest frequency o f occurrence o f 30 percent in the park co lle ctio n s. 9 10 The Aedes snowpool group was fu rth e r analyzed to species and species complex fo r each park c o lle c tio n . The results are shown in Fig. 2. The A- stimulans complex was most abundant in seven parks, comprising 100 percent o f the snowpool group in four park areas. The A- communis complex was present in samples submitted from three northern parks, being the most abundant in park 2 (73.15%). A* canadensis was the most abundant in park 3 where i t comprised 68.2 percent of the sampled Aedes snowpool group. The le a s t abundant of th is group was A cinereus in parks 2, 3, and 10 (1 2 .1 , 13.6, and 10.4%) and A. t r iv it t a t u s in park 1 (2.9%). Adult b itin g collections during 1972 and 1973 showed th at individuals o f the Aedes comminis- trichurus complex were the problem mosquitoes a t North Higgins Lake Park while an iden tical sampling program conducted a t Yankee Springs recreational area indicated that mosquitoes of the A* stimulans- f i t c h i i complex were responsible fo r b itin g problems there. Id e n tific a tio n of these collections was aided by keys contained in Barr (1958) and Carpenter and LaCasse (1955). Tables 2 and 3 show the results of the two-year study on b itin g a c tiv ­ ity a t the two parks. The highest a c tiv it y fo r A communis- trichurus occurred in la te May and early June with a mean count per s ite >,20 mosquitoes per 10 minutes. B iting a c tiv ity was reduced approximately 50 percent by the fourth sampling both years. A mean count of < 10 mosquitoes was recorded during the July 6 sampling in 1972 while a s im ilar level in 1973 occurred three weeks e a r lie r , on June 14. The b itin g a c tiv ity fo r th is species complex fo r both years was e s s e n tia lly n non-existent a fte r the th ird week in July. Other species id e n tifie d from the b itin g collection s from North Higgins Lake were A. canadensis, A_. tr is e r ia tu s , A. stimulans complex, A. vexans and C. perturbans. Of these only A., canadensis was consistently present in the collections beginning the f i r s t week in June and lastin g through the July c o lle c ­ tions. The mean b itin g count fo r th is species in 1972 ranged from 4.67 in the July 6 collections to 2.00 during the July 20 c o llectin g period while in 1973 the mean count ranged from 5.67 on July 5 to < 1 .0 0 on August 2. C. perturbans was collected in substantial numbers from only one s ite during 1972 and 1973. Individuals o f th is species appeared in the collection s during the f i r s t week o f July both years with the highest count o f 14 fo r a 10 minute sampling t r i a l recorded July 13, 1972 at the outdoor educational s ite . However, the species was never detected on a regular basis a t any o f the other s ite s . The other previously mentioned species were present so infrequently and in such small numbers that th e ir contribution to the b itin g problem was considered minimal. Table 3 shows the resu lts of the two-year sampling study at Yankee Springs recreational area fo r the A.. stimulans- f i t c h i i complex. The adult population had completely emerged by the end o f May 1972 and the mean b itin g a c tiv ity per s ite fo r seven weekly sampling t r ia ls ranged from 66.33 mosquitoes on July 4 to 51.00 mosquitoes on July 18. During the la s t week in July and f i r s t week in August the mean b itin g a c tiv ity per s ite dropped to 35.67 and 31.67, resp ectively. During and a fte r the second week in August the mean b itin g a c tiv ity per s ite 12 was < 10 mosquitoes. The b itin g a c tiv it y o f the A. stimulans- f it c h i i complex during 1973 was s im ila r. The average level o f b itin g a c tiv ity during the same time period ranged from (samples taken a day la te r from the 1972 t r i a l s ) 20.67 to 70.33. to August 2 was above 20.00. The mean b itin g a c tiv it y from July 19 During the sampling t r i a l s on August 9 and 16 the mean count dropped to below 10 per s ite . Other species id e n tifie d from the b itin g collection s were .A. canadensis, A_. vexans, A_. cinereus, A_. communis complex, £ . perturbans and Anopheles w alkeri. Of these, .A. canadensis and A. vexans were consistently collected in large numbers from the three public use s ite s during 1972 and 1973. Although A. canadensis was collected in sampling t r ia ls 1 through 78 both years the mean b itin g count fo r the three sites never exceeded 10 mosquitoes with the average b itin g a c tiv it y fo r 1972 and 1973 being 3.67 and 6 .5 2 , respectively. A^. vexans was present from the second week in June u n til the end of July both years with a high count of 8.33 recorded on June 28, 1973. Comparison of B iting A c tiv ity and E ffect on Recreation The level and duration of b itin g a c tiv ity o f the two species complexes were quite d iffe r e n t. Figs. 3 and 4 show the b itin g a c tiv ity fo r 1972 and 1973 at North Higgins Lake and Yankee Springs park areas, respectively. The ordinate scale measuring b itin g a c tiv ity is of the same magnitude in both graphs and represents an average to ta l o f mos­ quitoes collected fo r each sampling t r i a l during the two year study. 13 The average to ta l of mosquitoes collected a t North Higgins Lake during the f i r s t three t r i a l s was 74.50, 75.70 and 61.50, respectively. During the next three sampling t r i a ls mosquitoes collected averaged 26.50, 30.50, and 30.00 with the remaining samples in July averaging <20 mosquitoes. At Yankee Springs during peak b itin g a c tiv ity fo r A. stimulans- f i t c h i i complex, the average number collected ranged from 154.00 to 200.00 over a seven week sampling period (sampling t r ia ls 1-43). This was more than twice the average number collected a t peak biting levels and more than a twofold increase in the duration o f bitin g a c tiv ity as compared with the A.* communis- trichurus complex at North Higgins Lake. Whi.le the mosquitoes collected a t Yankee Springs aver­ aged 89.00 by the la s t week in July, th is level was s t i l l comparable to peak b itin g a c tiv ity at North Higgins Lake during la te May and early June. F in a lly the average number o f mosquitoes collected during the la s t two sampling t r i a l s (71 and 78) a t Yankee Springs was 27.00 and 18.50, respectively. Comparable values fo r the A_. communi s - t r i churus complex a t North Higgins Lake occurred approximately a month e a r lie r . Preliminary observations on the e ffe c t o f b itin g a c tiv ity on human recreation were also carried out. Fig. 3 shows that peak park attendance a t North Higgins Lake occurred during la te June, July and August when the mean b itin g count was < 10 mosquitoes per s ite . the mean b itin g collection s during la te May were ^ 2 0 , When park v is ito rs avoided recreational s ites located in forested regions ( i . e . , nature t r a i l s ) and stayed in those lo c a litie s th at possessed good prevailing winds (lake fro n ts ) and lacked a fo re s t canopy (beach areas). When 14 the mean b itin g counts were <_ 10 during June the recreational s ites in these forested areas were u tiliz e d by park v is ito rs . Recreational a c tiv itie s a t nature t r a i l and outdoor educational areas in Yankee Springs were adversely affected by the b itin g a c tiv ity of stimulans- f i t c h i i mosquitoes u n til the f i r s t week in August when mean counts dropped to £ 2 0 .0 0 mosquitoes. The b itin g annoyance at the nature t r a i l became so intense during la te June and July 1973 that the park n a tu ra lis t cancelled planned recreational a c tiv ity fo r th is area, mainly group tours fo r the purpose of nature studies. B iting collections made during th is time showed a mean count o f approximately 70 mosquitoes per s ite . A high o f 84 mosquitoes was collected in a 10 minute period at the nature t r a i l on June 28, 1973. Fig. 4 shows that park attendance and mosquito b itin g a c tiv ity peak e s s e n tia lly at the same time and a t levels much higher than a t North Higgins Lake. Dynamics of a Woodland Aedes Species Complex Table 4 shows the composition of the immature stages o f A.. s tim u la n s -fitc h ii at Yankee Springs in 1972. A ll collected on A p ril 12 were e ith e r f i r s t or second instars with the la t t e r representing 57.4 percent o f the to t a l. In the second co llec tio n 12 or 21.4 percent had matured to the th ird in s ta r while 30 were in the f i r s t in s ta r. A ll stages ( I - I V ) were present in the A pril 26 sample with fourth instars comprising 7.3 percent o f the to ta l and present fo r the f i r s t time. Immature development was rapid from A pril 26 to May 10. Of the 15 27 collected on May 3 a ll larvae were e ith e r th ird or fourth instars with the former comprising 62.9 percent o f the t o t a l. The la s t collection made on May 10 contained only fourth instars and pupal stages, each representing approximately 50 percent o f the to t a l. No f i r s t or second instars were present in the la s t two t r ia ls and adult emergence was f i r s t observed May 11 and continued fo r seven days. Egg hatch occurred e a r lie r in 1973 than in 1972 (Table 5 ). The f i r s t collection made on March 31 contained p rim a rily f i r s t and second instars although approximately 2 percent had progressed to the th ird instar. The number of f i r s t in s ta r present in the A p ril 3 sampling t r i a l increased su b s ta n tia lly (335 as compared with 94 the previous week) with 76.5 percent o f the to ta l being a t th is stage of develop­ ment. In the th ird t r i a l a ll larva l stages were present with 69.8 percent in the second, 22.9 percent in the th ird and approximately 2 percent in the fourth stage o f development. The fourth co lle c tio n on April 17 contained instars two through four with the m ajority (71.7%) s t i l l in the second stage. 5 percent. Fourth instars had increased to The la s t samples contained 76.3 percent fourth in s ta r larvae and 8.5 percent in the pupal stage. Adult emergence was f i r s t observed a week la te r on May 1 and continued fo r approximately 9 days. In addition, estimates were made on immature population densities in the aquatic microbasin area. The method used u tiliz e d the numbers collected and the size o f the hab itat area. Larvae and pupae counts were m ultiplied by a fa c to r obtained from the follow ing proportion to estimate the absolute density: to ta l surface a re a /to ta l surface area 16 sampled. The res u lta n t values are shown in parentheses below the actual sample values in Tables 4 and 5. Immature and Adult M o rta lity M o rta lity rates fo r immature stages were obtained from l i f e tables (Table 6) u t iliz in g data contained in Tables 4 and 5. During 1972 m o rta lity was estimated as increasing from 0.082 during the f i r s t age interval to 0.445 fo r the la s t. was 0.754. M o rta lity fo r the 28 day period Overall m o rtality in 1973 was estimated as 0.731 with no m o rtality estimated w ithin the f i r s t age in te rv a l. Rates o f 0.283, 0.179 and 0.543 were calculated fo r the remaining in te rv a ls . The substantial la te in s ta r m o rta lity observed during 1972 and 1973 was due larg ely to inadequate n u tritio n a l resources in the woodland pool breeding sites (Cummins).3 These depositional aquatic systems contained a v a rie ty of fin e p a rtic le d e tritu s u tiliz e d as food by mosquito larvae. As n u tritio n a l requirements increase s ig n ific a n tly fo r successive developmental stages, the amount o f a v a ila b le d e tritu s was seen as a lim itin g fa cto r. Although a chaoborid, Mochlonyx s p ., a predator of mosquito larvae, was consistently sampled throughout the two-year study, developmental rates of both insects were in phase and no s ig n ific a n t m o rta lity was observed. This suggested th at the A. s tim u la n s -fitc h ii ecosystem was e s s e n tia lly predator fre e during 1972 and 1973. Values fo r adu lt m o rta lity were obtained from l i f e tables (Table 7) u t iliz in g data from b itin g insect studies in 1972 and 1973 17 a t Yankee Springs recreational area. As the surveys were conducted beneath a dense fo rest canopy, adverse weather conditions had minimal e ffe c t on b itin g a c tiv ity . S im ilar b itin g counts fo r A_. stimulans- f it c h i i at d iffe re n t c o lle ctio n s ites indicated the mosquitoes were regularly d is trib u te d throughout the park. Research (Detinova, 1968) has also shown th at u n iv o ltin e , e a rly season Aedes mosquitoes surviving to August completed 6 gonotrophic cycles ind icatin g a requirement fo r successive blood meals during the summer months. For these reasons, substantial departures from peak b itin g levels were considered a res u lt of m o rta lity factors a ffe c tin g the mosquito population. Conversely, m o rtality was considered minimal a t peak b itin g levels (age in terva ls 0 -36). L ife tables fo r adu lt A. s tim u la n s -fitc h ii show th a t a popula­ tion cohort is long liv e d with 0.12 surviving to age in te rv a l 78-85 in 1972. During the same time period in 1973, 0.059 were estimated as surviving. In terms o f actual time o f ye ar, th is age in terval represents the second week in August. Since th is species complex is univoltine and adult emergence was complete by May 31, females sur­ viving to the la s t age in te rv a l were approximately three months old. Zero m o rta lity was assumed fo r one month a fte r emergence. Empirical observations indicated th a t adult males were present fo r approximately two weeks a fte r i n i t i a l female emergence. Footnotes Jou rn al a r t ic le no. 6906, Michigan A g ricultural Experiment Station. This research was supported by a fellow ship grant from the Michigan State Department o f Natural Resources. 2 V. Wagner. Data from a completely randomized transect sampling procedure was analyzed by a nested analysis o f variance. The variance component due to pools was not s ig n ific a n t a t the .05 or . 0 1 levels (unpublished data). 3 K. Cummins. Personal communication, September 1973. Table 1. Park Percent composition of the major mosquito species in landing/biting collections made in ten state parks, 1971 Total No. Aedes snowpool group C o q u ille ttid ia perturbans Aedes vexans Anopheles quadrimaculatus Aedes tris e ria tu s Culex pipiens 1 70 50.0 32.9 1.4 — 7.1 7.1 2 367 90.2 1 .1 7.4 — 1.3 — — — — — - - — — 3 28 4a — 5 1 12 8 8 . 0 1 2 . 0 — — 30.4 <1 . 0 10.7 46.4 — — 6 . 2 28.4 4.9 — — — — 23.3 6 81 58.0 2.5 7 21 80.9 19.1 — 8 30 40.0 -- -- 9 53 — 92.3 — 10 69 42.0 20.3 11 17 23.5 76.5 23.2 — aNo adult samples submitted from North Higgins Lake. — 30.0 7.7 — — 1 0 .1 - - — — 4.4 — Table 2. Biting a c tiv ity of Aedes communis-trichurus complex at North Higgins Lake state park, 1972-1973 Month/Daya 1972 May June July Sampling T ria l Sample Total x Biting Count/10 Min Sx 1973 1 Week Interval 1972 1973 1972 1973 1972 1973 25 24 1 81 6 8 27.00 22.67 2.646 2.186 31 31 8 8 6 65 28.67 21.67 2.404 1.732 15 64 59 21.33 19.67 1.764 2.028 8 . 0 0 1.453 1 .0 0 0 8 7 15 14 22 29 24 9.67 22 21 29 35 26 11.67 8.67 1.856 1.453 29 28 36 31 29 10.33 9.67 0.882 1.856 5 43 19 16 6.33 5.33 0.333 0.333 1 0 9.33 3.33 0.882 0.882 — 1.67 — 0.667 6 13 12 50 28 2 0 19 57 5 — C o lle c tio n s made a t weekly intervals from time of adult emergence u n til mean biting count per s ite was < 1 0 / 1 0 min. period fo r two successive weeks. Table 3. Biting a c tiv ity o f Aedes stimulans- f it c h i i complex at Yankee Springs recreational area, 1972-1973 Month/Daya 1972 May Sampling T ria l 1973 1 Week Interval 31 31 6 7 June Sample Total x Biting Count/10 Min Sx 1972 1973 1972 1973 1972 1973 1 168 141 56.00 47.00 4.163 10.017 8 181 175 60.33 58.33 4.631 8 .2 1 2 13 14 15 190 199 63.33 68.67 3.283 7.311 2 0 21 22 189 195 63.00 65.00 3.055 7.056 27 28 29 185 211 61.67 70.33 3.844 8.413 4 5 36 199 201 66.33 67.00 4.177 7.371 July 11 12 43 176 157 58.67 52.33 8 . 1 1 0 18 19 50 153 62 51.00 20.67 6.028 2.603 25 26 57 107 71 35.67 23.67 8.090 2.667 2 1 . 0 0 7.688 5.132 Aug. 1 2 64 95 63 31.67 8 9 71 29 25 9.67 8.33 0.667 2.028 16 78 25 12 8.33 4.00 0.882 1 .0 0 0 15 was < 10.493 C o lle c tio n s made at weekly intervals from time of adult emergence u n til mean biting per s ite min. period fo r two successive weeks. 1 0 /1 0 Table 4. Immature Aedes s tim u la n s -fitc h ii in a woodland pool at Yankee Sprinqs recreational area, ! 972 No. of Larval Instars ( I- IV ) & Pupae (P) Date 4/12 Sampling T ria l I 1 2 6 (22,515) 4/19 4/26 5/3 5/10 2 3 4 5 30 (25,979) 14 (12,123) II a III IV P 35 (30,308) Total 61 (52,823) 14 (12,123) (10,391) 12 12 (10,391) (10,391) 56 (48,493) 12 17 (14,721) 3 (2,598) 41 (35,503) 27 (23,381) 10 (8,660) 8 (6,928) Estim ated to ta l number of larvae in woodland pool in parenthesis. bAdult emergence f i r s t observed on 5/11/72 and completed 5/17/72. 7b (6,062) 15 (12,990) Table 5. Immature Aedes s tim u la n s -fitc h ii in a woodland pool at Yankee Sprinqs recreational area, 1973 No. of Larval Instars ( I- IV ) & Pupae (P) Date 3/27 Sampling T ria l 1 4/3 2 4/10 3 4/17 4/24 4 5 II 94 (27,166)a 41 (11,849) (578) 137 (39,593) 335 (96,815) 94 (27,166) 9 (2,601) 438 (126,582) 17 (4,913) 219 (63,291) 72 (20,808) (1,734) 314 (90,746) 60 (53,465) 13 (17,340) (3,757) 258 (74,562) 18 (5,202) 90 (26,010) 185 III IV P I 2 6 Estim ated to ta l number of immature stages in woodland pool. bAdult emergence f i r s t observed on 5/1/73 and completed by 5/10/73. 1 0 b (2,890) Total 118 (34,102) Table 6. L ife tables fo r immature Aedes s tim u la n s -fitc h ii in a woodland pool ecosystem at Yankee Springs recreational area, 1972-1973a Age Interval (Days) No. Surviving at S ta rt of x X 0 X 1 ,0 0 0 No. Dying Within Interval x to x + 1 d M o rta lity Rate q b X 1 ,0 0 0 82 1 ,0 0 0 246 Survival Rate p c Kx 0.082 0 .0 0 0 0.918 1 .0 0 0 283 0.268 0.283 0.732 0.717 0 7 918 14 672 717 229 128 0.341 0.179 0.659 0.821 21 443 589 197 269 __ 0.543 _- 0.457 246 0.445 •— 0.555 28 320 —— aThe two coluntis under each heading are fo r the years 1972 and 1973, respectively. ^Overall m ortality rates fo r 1972 and 1973 are 0.754 and 0.731, respectively c0verall survival rates fo r 1972 and 1973 are 0.246 and 0.269, respectively. y . .. .... Table 7. L ife tables fo r adult Aedes stim u la n s -fitc h ii at Yankee Sprinqs recreational area, 19721973a Age Interval (Days) No. Dying Within Interval x to x + 1 No. Surviving a t S ta rt of x X Survival Rate M o rta lity Rate Px dx 0-36 1 ,0 0 0 1 ,0 0 0 36 1 ,0 0 0 1 ,0 0 0 0 0 0 .0 0 0 0 . 0 0 0 1 .0 0 0 1 .0 0 0 156 219 0.156 0.219 0.884 0.781 473 0.131 0.606 0.869 0.394 0 0.301 0 . 0 0 0 0.699 1 .0 0 0 0 0 .1 1 2 0 . 0 0 0 0 .8 8 8 1 .0 0 0 43 844 781 1 1 0 50 734 308 221 57 513 308 58 64 455 308 316 186 0.695 0.603 0.305 0.397 71 139 122 19 63 0.138 0.520 0.862 0.480 78 1 2 0 59 1 .0 0 0 1 .0 0 0 0 . 0 0 0 0 .0 0 0 85 0 59 0 1 2 0 — — — — — — aThe two columns contained under each heading are fo r the years 1972 and 1973, respectively. po cn 26 Fig. 1. Park and recreational areas in Michigan where mosquito surveys were conducted during 1971. is as follow s: Size in acres o f each park 1— 58,327.2; 2— 19,244.5; 3— 9,1 38 .0 ; 4—428.0; 5 - 1 9 6 .3 ; 6— 963.0; 7— 981.0; 8— 9,612.6; 9— 17,053.2; 1 0 -4 ,9 7 1 .6 ; 11—4 ,156.6. 27 O n ton og on fla ro g o G o g a b ie ” ' ^ MorouaM a Alg tr D ic im a o n Cmppawo : S e ho olcroM r Moekm'o 'c D tllo Em m tt Chaboygon C nortavon P rta qu a O U a g o M o n lm o ra n c y j Craw ford vKolkoaho Banna Porcupine Mtns. G rand Trav arao W a if or d • Oa codo i»ia A lp tno Alco no • jM ia ao uka t a c om m on j O g am aw loaeo Tahquamenon F a l ls H a rtw ic k Pines 4. Higgins Lake (North) 5. Bay C i t y M aa on | Lo ka O a c ao lo • Nawoygo Mae o ■ t a C lara A rtn oc 11 Oeaono M id la n d i Sleeper Tui colo 7. Aigonac 8. Pinckney M o n lc o lm G ro iio l Sogm ow Waterloo Ganaaat Lopaar Lud i ngton Bo r r y Alla g o n van B uran | Si 01 f o w o Yankee Springs -Soniiac E a to n K o l a m o t o o ; Calh oun St Joaaph;Q ronen Figure 1 in g fia m Jaekaon H i ii a d o ia - - ~r ; u v m g a t on O a ila n d wayna M onroe C lo 28 Fig. 2. Percent species composition of the Aedes snowpool mosquitoes id e n tifie d from collections made in nine Michigan state park and recreational areas, 1971. A . STIM U LAN S COMPLEX A. COMMUNIS COMPLEX A. CANADENSIS A . C IN E R E U S A . T R IV IT T A T U S 100.0- S 3 PERCENT SPECIES COMPOSITION 9 0 .0 80 .0 7 0 .0 6 0 .0 5 0 .0 4 0 .0 30 . 0 20 .0 - . - 10 0 PARK Figure 2 8 1 0 30 Fig. 3. Average mosquito b itin g a c tiv ity and park attendance fo r 1972 and 1973 a t North Higgins Lake state park. MOSQUITO BITING A CTIVITY PARK ATTENDANCE -24 22 20 € m m 18 * 16 14 > 30 7i $ -I 12 10 *\ \ / V m o > o m 8 c co > o (0 JUNE 1 JULY Figure 3 AUG. I SEPT. 32 Fig. 4. Average mosquito b itin g a c tiv ity and park attendance fo r 1972 and 1973 a t Yankee Springs recreational area. t MOSQUITO BITING ACTIVITY 220 55 200 50 180 45 160 40 140 35 1 2 0 30 ATTENDANCE 1 0 0 25 80 20 60 -115 40 10 20 5 MAY I JUNE JULY Figure 4 AUG. SEPT. IN THOUSANDS TOTAL/I0 -.60 PARK SAMPLE 240 WEEKLY MIN. TRIAL PARK ATTENDANCE PART I I A SYSTEMS SCIENCE APPROACH TO THE STUDY OF BITING INSECT POPULATIONS IN MICHIGAN STATE PARKS1 INTRODUCTION A previous survey has established th at species o f the A. stimulans and A. communis complexes are la rg e ly responsible fo r b itin g insect problems in Michigan state parks. These mosquitoes are the woodland v a rie ty and u t iliz e temporary le n tic pools as breeding s ite s . These pools are formed as a re s u lt o f snow melt and flooding conditions prevalent in forested areas during e a rly spring. Due to the importance of the woodland Aedes group as potential disease vectors and nuisances in the park system, a study was undertaken to c le a rly define the r e la ­ tionship between the aquatic environment and the immature l i f e stages. A systems science approach was used to provide an operational framework to determine what biological and aquatic features were needed fo r th is study. Additional benefits th a t would be gained from th is approach are id e n tific a tio n of sp e c ific data requirements and research p r io r itie s necessary in a w ell-coordinated biological study. The f i r s t step in constructing th is systems model was selecting model components and adequate state variables to represent them, a process known as aggregation (Patten, 1972). The Aedes l i f e cycle and the respective aquatic breeding s ites are viewed, fo r the purposes of th is study, as a co llection o f in te ra ctin g components. When one set of components is chosen fo r a p a rtic u la r model, the robustness o f the real system is lo s t. This was the d if f ic u lt y encountered when the 35 36 i n i t i a l choice o f components and behavioral features fo r the Aedes ecosystem was made. As a r e s u lt, relian ce on previous biological information and data was necessary in structuring the Aedes system model to obtain a reasonable facsim ile o f the natural ecosystem. DEFINITION OF THE AEDES ECOSYSTEM The modeling process was concerned with those univoltine Aedes mosquitoes th at u t i l i z e temporary le n tic pools as developmental sites fo r the immature l i f e stages. Upon emergence from these aquatic areas, the adults assume a te r r e s tr ia l existence fo r the remainder of th e ir l i f e span. Since a system may be defined as a collection of objects, each behaving in such a way as to maintain behavioral consistency with its environment as w ell as objects in the system, the l i f e cycle and environmental interactions o f this group o f Aedes mosquitoes were p artition ed into a conceptualized ecosystem (Caswell e t a ! . , 1972). As a re s u lt the various phases of mosquito development became a l i f e system with in s ta r and adult stages represented by sp ecific components. The expression of th is phenomenon as a structured diagram is shown in Fig. 1. E s s e n tia lly , i t represents a population flow chart with l i f e components coupled to physical components representing the woodland pool! . 2 Components th at in te ra c t exclusively with the woodland pool are the spring egg and immature ^ stages while the adult Vg and summer egg Yy stages are placed in a position of interaction with the te r r e s tr ia l environment. 3 Those components representing the aquatic environment are the dissolved oxygen concentration ¥-j, water temperature ^ and water depth of the woodland pool. They were chosen from an in f in it e array of aquatic parameters as key factors in the response of the spring egg and immature stages. 37 38 Because a physiological reorganization occurs between the summer egg ¥ 7 and spring egg ¥ 4 components and no attempt was made to model these physiological states (diapause, conditioning, re a c tiv a tio n ), a break designated as overwintering delay is positioned a t th is point in the system. During th is time a changed physiological state and environment are mandatory fo r the continued functioning of the system. The two Vg 5 components designating the pupal stage serve to separate the population according to sex. Since fin a l adult d iffe re n tia tio n is realized during th is stage, a d e fin itiv e sex r a tio resu lts at the time o f emergence. That portion of population designated as the adult female component Tgp w ill have d ire c t input into the aquatic environment while the adult male component is perceived as an end point in the system. The next step in th is sequential structuring process was to describe the behavior o f the components. A behavioral description specifies an o b ject's overall behavior through time in terms of behav­ io ral features selected by the observer. At any point in tim e, the behavior of a component in the Aedes system is specified as the values assumed by a set of behavioral features ( i . e . , stim uli and responses). The choice o f behavioral ch aracteristics as well as the components o f the Aedes ecosystem was i n i t i a l l y made from a vast array o f features on the basis o f knowledge accumulated by the author. Construction of the model was aided by a review of entomological journals supplemented by the studies conducted in Michigan state parks. Since the primary goal was to study immature-aquatic in te ra c tio n , emphasis was placed on 39 those behavioral features th a t described the behavior of th is portion of the ecosystem. With these concepts taken into consideration, the diagram shown in Fig. 1 was broken down into its separate components fo r the purpose o f individual study. This set and th e ir respective behavioral features are shown in Table 1. O rientation o f Behavioral Features into a Stimulus and Response Set Once the set o f behavioral features was selected, they were oriented into a stimulus-response set. By jo in in g the components at th e ir points of in teraction a system graph was constructed (Fig. 2 ). At the points o f interconnection the behavioral features are perceived as responses (determined w ith in a component) or stim uli (determined at the interconnection or outside the system). These two configurations are represented in Fig. 2 by the general form notation Rb^ and Sb^, respectively. The behavioral features oriented as stim uli to the aquatic environment components - Vg are (with the exception of water temp Sb^) determined outside the system and were viewed as external stim uli to the Aedes system. The aquatic component ^ representing dissolved oxygen interacts exclusively with the spring egg stage ^ while the components fo r water temperature and le v e l, ^ anc* ^3 * in te ra c t with the immature l i f e stages, V,- .. I t should be re 3 »J emphasized th at because a precise description o f aquatic-immature interaction was required, a component 4V . was needed to represent each of the fiv e stages of the immature portion o f the l i f e cycle. 40 To view the immature component as having no intern al structure resulted in the in v a lid assumption th a t the aquatic stim uli a ffe c t a ll immature stages equally. The response variables o f the spring egg and la rv al stages were a function o f the stim uli they receive from the aquatic components and o f th e ir state variables. In Fig. 2, the populations in various l i f e stages were designated as state variables and adequately described the intern al structure o f the population. Since we were interested prim arily in control o f aquatic l i f e stages and since the adu lt and summer egg stage were interacting only with the te r r e s tr ia l environment, we did not model in d e ta il the behavior of these l i f e stages. While the occurrence o f te r r e s tr ia l interactions cannot be questioned, these components were not seen as in te ra c tin g with the aquatic environment, and the exclusion o f external t e r r e s tr ia l stim uli w ill not a ffe c t the adequacy o f the proposed simu­ la tio n . In representing the intern al behavior of egg input by the adult Vgp stage, R b^ was viewed as occurring a t 24 one-inch levels along the sides of the pool. These individual gradients are designated as the area in which eggs are positioned fo r eventual spring flooding, controlled by the hatching stimulus Sbg (water depth). egg ¥ 7 and spring egg ^ a column vector. Thus the summer are viewed as sets o f 24 variables arranged as 41 Construction o f Free Body Models The next step in th is approach was the construction o f free body models in is o la tio n from other parts o f the system. The breaking of the system's components into free body form was required due to the sheer complexity of the Aedes ecosystem. In essence, free body models p ro h ib it making the behavior o f one component a d ire c t function of the behavior o f another component. The free body model of a component is a set of functions which specifies the values fo r the responses and state variables based on the stim uli presented and the state of the system (which retains information about the stimulus history of the o b je c t). Fig. 3 diagrams the components o f the Aedes system in free body form and Table 2 lis t s the free body equation forms of these models. The state and response equations take the follow ing general form. - F, ( * , . S ,. t ) Ri = 6i ^ i * Si ’ ^ (Caswell e t a l . , 1972) The free body models fo r the f i r s t in s ta r and pupal stages were con­ structed separately as each possesses a behavioral feature (maturation to f i r s t in s ta r Sb-jQ and emergence to adult R b ^ ) not possessed by the others. The rest o f the immature stages 4 3ij ( j = 2, 3, 4) have iden- tic a l behavioral sets and are modeled id e n tic a lly (but with d iffe re n t parameter values). As component 4 ^ was viewed as an end point in the system, no response equation is generated. I t is fo r the simultaneous 42 solution o f series of equations o f th is general form that a computer simulation o f the Aedes model was constructed. A discussion of actual free body model equations used in the computer program is beyond the scope o f th is paper. However, two examples are b r ie fly described in the discrete-tim e form used in the simulation. The dynamics of the component water depth ^ is expressed and updated in the following state and response equations: Y3 (t+ 1 ) = [¥ 3 ( t ) + (Sb2 )(Sbg)(AREAF) - Sb?] [0.95 +0.05 (Sbg) ] Rbg(t) = H' 3 ( t ) where Sbg = 0 .9 - 0 .0 1 (DAY). Runoff fra c tio n Sbg into the pool was computed from a lin e a r approximation of data obtained from the National Weather Service fo r average runoff proportion ( i . e . , 1 -proportion absorbed) in Michigan during the months o f March through June. It was m u ltip lie d by an area facto r (AREAF) of 10 to represent approximate ra tio of microbasin area to surface area o f the pool .* (Currently a time lag of one day fo r 0.5 of the runoff from a day's p re c ip ita tio n has been added to the computer sim u latio n.) The term [ 0 .9 5 + 0.05(SBg)] represents the calculation o f the proportion o f water not absorbed into the ground and was our rough estim ate, pending resu lts of subsequent studies. P re c ip ita tio n Sbg and evaporation Sby are d a ily inputs into the simulation model and were obtained from data supplied by the National Weather Service at Michigan State U niversity. The biological components were much more d if f ic u lt to model due to th e ir complexity and lack o f information on many of the specific 43 behavioral mechanisms. Therefore results o f experimental work conducted on Aedes physiology were i n i t i a l l y incorporated into portions of the simulation equations. component The l i f e stage discussed is the spring egg a vector of 24 one-inch levels representing egg position along the flood lin e o f a woodland pool. Since a break in the system occurs between components 4^ and 4^, a tran sfer of la s t year's viable eggs from Ty to is accomplished by the simulation a t the beginning of the current season. Whenever the water temperature Sb^ is > 5° C and dissolved oxygen concentration is < 8 ppm hatching may occur. Other conditions fo r egg hatch are: FDO > PDO hatch = 0 FDO = PDO continue started hatch but do not in it ia t e new hatch FDO < PDO hatch and i n it i a t e new hatch continue started where (FDO) and (PDO) are the current and previous day's dissolved oxygen concentration, respectively. When hatching is determined to be possible, the state equation fo r each submerged gradient level i ( i . e . , i < Sbg+ 0.5 where the addition of 0.5 allows a gradient oneh a lf to completely flooded to be designated as submerged) must be recalculated as follows: V4 . l ( t + 1) = ^ 4 , i ( t ) ( 0 * 5 ) To simulate to the delayed hatching e ffe c t, h a lf o f the eggs are carried over next day while the remainder e ith e r hatch or become non-viable according to the response equation: 44 Rb1 0 ,i ( t ) = C*4,1 ( t ) ] C {1 -0.1225(S b3) ] 0.5 The term [1 -0 .1225 ( S b g ) ] is a lin e a r approximation expressing the percentage of eggs th at w ill hatch as a function o f dissolved oxygen Sbg (Judson, 1960). When hatching conditions are not met the state and response equations take the following form: * 4(t + l ) = ? 4 (t) Rb1 0 ( t ) = 0 The to ta l egg hatch in the woodland pool is determined by summing Rb^ .j(t) over a ll submerged egg containing gradients, i . e . , i f N6 = Min (24, Sbg+ 0 . 5 ) Rb] 0 NG (t) = ^ Rb1 0 > 1 (t) These, then, are examples of actual free body equations constructed fo r two components of the Aedes system. The procedure is repeated fo r a ll components (V-j, Y2, and ^ g - V y ) with the res u lta n t equations being u tiliz e d in the computer simulation model to describe the dynamics of the Aedes l i f e system. 45 Footnotes Journal A rtic le No. 6804 Michigan A gricultural Experiment S tation. This research was supported by a fellowship grant from the Michigan State Department of Natural Resources. Associated with each component i is a state variable The variable ¥ -j(t) describes the state o f the component a t time t , and it s value is continually updated as a function o f the component's state and stim uli a t time t . The state variable often represents an accumulation o f m a te ria l, energy, or organisms a t some point in the system, and instances of each of these uses are found in th is model. 3The determination o f immature-aquatic interactions can be deduced on the basis of a single l i f e component ¥ 5 as a ll immature stages are in teractin g with id e n tica l aquatic features. However, the responses to these features are not id en tical and i t was neces­ sary to generate fiv e d is tin c t state and response equations. Hence the ¥5 , j ( j = 1, 2, 3, 4 ,5 ) designation was used to indicate in s ta r and pupal responses. ‘‘ (Variable names) used throughout th is portion of the paper are the actual variable names u tiliz e d in the computer program. 46 Table 1. Behavioral features associated with components of Aedes ecosystem with emphasis on immature-aquatic interactions Component Ti Behavioral Features Dissolved Oxygen HgO Temp. bi b2 Precip. CM o o b3 * 2 Water Temperature HgO Temp. bl A ir Temp. (Max. & M in.) b4 Radiation »S Soil Temp. b 6 *3 Water Depth Precip. b? Evap. b7 H20 Depth b 8 Runoff b9 *4 Spring Egg Stage HgO Temp. bl do2 b3 H20 Level b 8 Maturation to 1st Instar bin Egg Overwinter b14 *5.J Immature Stages j = 1, 2, 3, 4, 5 11 b 12 Female Adult Stage b 1 2 b13 'f 6 M *7 Male Adult Stage Summer Egg Stage H20 Level b 8 bio b-i*. % H20 Temp. bl b 12 hs b14 Maturation to 1st Instar j »j Maturation from In star j to (j + 1 ) Emergence to Adult Emergence to Adult Egg Lay Emergence to Adult Egg Lay Egg Overwinter Table 2. State and response equation forms generated by the free body models of each component in the Aedes system State Equation Forms = F(¥r ^ Sbr Response Equation Forms Sb2 , t ) Rb3 = , Sb] , Sb2 , t ) Y2 = F(V2# Sb4> Sb5 , Sb6 , t ) Rb1 = G(Y2 , Sb4 , Sbg, Sbg, t ) = F(¥3, Sb2 , Sb?, Sbg, t ) Rbg = G(Y3, Sb2 , Sb7, Sbg, t ) ^ ■ ¥ 3 H ¥4 t 1 = F^ 4 ,1 » Sbl* Sb3 ’ Sb8 * Sbl 4 ,i * t ) i = 1 , 2 , . . . 24 = F^ 5 ,r A ¥5 , j = F^ 5 , j * Sbr cfif ^5 , 1 Sbr Rb-jQ = G(,F4 j , Sbi ’ ^b3 * ^b8 * ^b14 i * ^ i = 1, 2, 24 Sb8 * Sbi o ’ ^ Rbl l , l = S^ 5 , l 5 Sbl* Sb8 ’ Sb1 0 * ^ Sb8 * Sbn , j - r Rbn , j = G^ 5 , j 5 sbr j = 2, 3, 4 *cRf ^5,5F = F^ 5 , 5 5 cfif V Sbr Sb8 * Sbl 1,4 s ^ = F^ 6 F* Sb1 2 ’ ^ Sb8 * Sbn , j - r 3 = 2, 3, 4 Rbi 2 = G^ 5 , 5 ’ Sbl ’ Sb8 ’ Sbl l , 4 ’ ^ Rbl 3 = G^ 6 F’ Gbl 2 ’ dt V6 M = F^ 6 M’ Sb1 2 ’ H ^ 7 , i = F^ 7 ,i> Sbl 3 ,i * *> i = Rbl 4 , i = G^ 7 , i 5 Sbl 3 ,i * ^ 1 , 2 , . . . 24 id e n tic a l equation fo r ^g ^ i = 1, 2, . . . 24 48 Fig. 1. A diagram o f woodland Aedes population components and th e ir respective environments. Components ^ , ¥2> V3 , represent dissolved oxygen concentration, water temperature, and water depth of woodland pool, resp ectively. Component spring egg stage. Component ( j = 1, 2 , 3, 4 , 5 ); 4 la rv a l instars and pupil stages where ^ ¥ 5 5 j is subdivided into a male and female set. Component adult stage subdivided into a male and female set. Component summer egg stage. OVERWINTERING \ DELAY I------- TERRESTRIAL CONDITIONS % (V. AQUATIC CONDITIONS Figure 1 LIFE CYCLE 50 Fig. 2. System graph of Aedes mosquito ecosystem indicating stimulus Sb.. and response Rb^ o rie n ta tio n of the behavioral features. Q Sb7 —O -O - Rb, 6M 6F Rb| Sb Rb Sb, 11.2 Rb, Rb, 11,1 Sb Sb, Sb, Rb, Sb, Sb, 5,2 11.3 5.3 Rb, 11.3 5,4 11.4 Sb, Sb, Rb, Sb Rb Sb, Sb, Figure 2 Sb, 52 Fig. 3. components. Construction o f free body models associated with the Aedes Sb Sb Rb Sb. Rb u.j -O---- -O- — - O n.J-l ■ i Sb Sb. Rb Sbe Rb Sb. O— -O Sb,. ■O- O Sb, i i cn Sb CO Sb Sb -a o- Rb, -O 5,5 F ii,4 'IDENTICAL FREE BODY DIAGRAM fo r% Rb ■6M ,5m Sb, -a Sb, Figure 3 I4 J * 1 1.2 -24 1*1,2, -24 PART I I I CONSTRUCTION AND REFINEMENT OF A COMPUTER SIMULATION MODEL FOR POPULATION STUDIES OF WOODLAND POOL AEDES MOSQUITOES1 INTRODUCTION Once a system model fo r Aedes woodland mosquitoes was obtained, a computer simulation was constructed to simulate the dynamics of the insect and of the woodland pool ecosystem. A computer simulation is merely one of the methods av ailab le fo r examining behavioral sequences and external stim uli to an ecosystem. In some instances th is approach to biological studies can y ie ld very powerful resu lts (Caswell, 1972). However, the inherent problem in modeling ecological systems is that the assumptions necessary to achieve an a n a ly tic a lly acceptable struc­ ture may be so gross that the resu lts o f analysis have minimum relevance to the real world. Therefore constant experimentation with the Aedes ecosystem and comparison with inherent q u a litie s and defects o f the model must be undertaken by the b io lo g is t to insure a representative simulation. The optimum approach may be a combination of an a lytic al techniques ranging from computer simulations to numerical solution and graphical analysis. Comparison of resu lts would give an approximation of the model's s e n s itiv ity in describing real world situ a tio n s. Equations used in the computer program were formulated with the aid of a system graph developed to study th is entomological problem. Information and data u tiliz e d in the model construction were obtained from surveys conducted in Michigan state parks and an extensive lite r a tu r e review. 55 MODELING THE BIOLOGICAL COMPONENTS As in any modeling process assumptions were made before the actual model construction began. Following are the reasoning processes underlying the equation forms used in the sim ulation. I t was assumed that water temperature was the key fa c to r in determining the maturation rates of immature stages. Of the various attempts to mathematically describe water temperature e ffe c t on mosquito larva l growth rates (H uffaker, 1944; L a i, 1953; Haufe and Burgess, 1956; Nielsen and Evans, 1960), the catenary formula (Janiscfi, 1932) u tiliz e d by Huffaker (1944) was chosen fo r use in our sim ulation. The choice o f th is formula met the requirement of precise mathematical descriptions fo r individual immature stage growth rates. This was necessary because each in s ta r and pupal stage growth rate response was d iffe r e n t and required the generation of fiv e d is tin c t state and response equations. Also, the addition of a p o sitive and negative exponential curve used in th is formula (representing the acceleration o f maturation rates through ris in g temperature and retard atio n during the developmental process, resp ectively) is preferred to expressing developmental rates as a lin e a r function o f temperature since a s tra ig h t lin e relation ship does not adequately explain insect development a t the extremes o f the temperature range. The actual progression o f immature cohorts through the la rv a l stages was assumed to be uniform with minimal variance of growth rates w ith in a cohort. 56 57 Assumptions concerning the dynamics of egg hatch took into consideration the ro le o f dissolved oxygen concentration in in it ia tin g the egg hatch response. Numerous authors have studied this phenomenon (Borg and H o rs fa ll, 1953; H orsfall e t a l . , 1958; Judson, 1960; Trpis and H o rs fa lls 1967) and have shown a decreasing gradient o f dissolved oxygen was required to in it ia t e the egg hatch response. A lin e a r approximation o f data (Judson, 1960) concerning the e ffe c t o f s ta tic levels of dissolved oxygen on two species o f Aedes mosquitoes was i n i t i a l l y u tiliz e d in the sim ulation. The data were obtained from that portion o f the experiment u t iliz in g eggs stored at 90-100 percent re la tiv e humidity since the assumption was th a t eggs in a natural environment are positioned in locations th at maintain a high soil moisture gradient. In the park study areas these locations are moist layers of plan t d e tritu s along the flood lin e o f woodland pool micro­ basins and are usually protected from moisture loss by a dense fo rest canopy. The storage of eggs in the top 24 inches of a pool area was viewed as the egg input response by the adu lt female population. The females are assumed to oviposit the eggs in the gradients closest to the receding water level and the point o f maximum flood. This assump­ tion agreed with H o rs fa ll's (1963) observations concerning egg po sitio n­ ing o f Pl stimulans mosquitoes in woodland pool microbasins. A ll surviving ad u lt females were assumed to have taken a blood meal needed fo r egg formation since the park areas support substantial populations of small mammals and are v is ite d by thousands o f campers during the summer months. The gonotrophic cycle o f the woodland Aedes mosquito 58 was assumed to occur over a seven day period with egg oviposition occurring every seventh day with each female laying an average of 70 eggs along the flo o d lin e of the woodland microbasin. These assumptions were based on observations (B arr, 1958) concerning the Aedes stimulans complex indicating th at the preoviposition period was approximately seven days with a mean egg lay o f 70 per female. The results o f that study also indicated no appreciable delay in o v i­ position by females of this species complex. F in a lly , f e r t iliz a t io n of the emerging females was assumed to be 100 percent. Once these assumptions were made and incorporated into the modeling process, a computer simulation of the Aedes system was attempted. Dynamics of the Immature Stages The most important biological component from the aspect of the modeling process was the one representing the immature l i f e stages. The dynamics of movements of organisms w ithin and among these stages are made by the program on a d a ily basis and controlled by three subroutines and a function subprogram (Fig. 1A). The immature stages are represented in the program by a twodimensional vector IN ( i , j ) . 2 The f i r s t subscript refers to the number of days a population cohort has been present in a p a rtic u la r stage, while the second subscript refers to the various stages as follow s: 59 a value o f refers to 1 in s ta r 1 2 in s ta r 2 3 in s ta r 3 4 in s ta r 4 5 male pupa 6 female pupa For example, IN (1 2 ,2 ) contains a count of the number of larvae which have been in in s ta r 2 fo r 12 days. Subroutine development w ithin each o f these l i f e stages. DEVLOP d irects the Every day, the elements of each stage are advanced w ithin th at stage to the next p o sitio n , subject to a d a ily m o rta lity rate (stored as a vector, K IL L (j), where j takes values according to the table above). A 20 day lim it has been selected as a maximum expected duration fo r any one stage; however, in theevent that a longer progression is required, cohortssimply in the 2 0 th element (once reached) u n til graduation to the remain next stage. This process is expressed by the following equations ( IN (2 0 ,j) + I N ( 1 9 , j ) ) * ( l - K IL L (j)) IN (i,j ) = ( 0 I N ( i . j ) * ( l - K IL L (j)) i f i = 20 if i = 1 otherwise where i sequences in u n it decrements from twenty to one, and j sequences in u n it increments from one through s ix. Subroutine GRADUAT graduates an immature cohort (group of individuals o f the same m aturity a t a given time) to the next in star 60 stage dependent on the rate o f development calculated by subroutine GRADRAT. The l a t t e r subroutine u tiliz e s a function subprogram CTNRY to set RATE(I); the number o f days of development required before movement from one immature stage to the next can be accomplished. Function CTNRY u t iliz e s the catenary formula which describes the duration o f each la rv a l and pupal stage and takes the following form t = where t = time, m = developmental time at the em p iric ally determined optimum, a = constant determining the slope of the curve, and T is the temperature in degrees above or below the optimum. Then, each day, subject to certain co n strain ts, immature cohorts in each stage may be graduated into the succeeding stage. The f i r s t in s ta r stage must be presented with a water temperature in excess o f 8 ° C p rio r to graduation to the second in s ta r; while the second in s ta r requires a 15° C minimum temperature before graduating to th ird in s ta r. The equation below d e ta ils the accumulation of cohorts fo r graduation; while the table below summarizes the graduation mechanics: 20 i=RATE(j) 61 cohort represented by th is to ta l is added to this location SUM fo r stage 1 IN (1 ,2 ) second in s ta r, day 1 SUM fo r stage 2 IN (1 ,3 ) th ird in s ta r, day 1 SUM fo r stage 3 IN (1 ,4 ) fourth in s ta r, day 1 J5 SUM fo r stage 4 IN (1 ,5 ) male pupa, day 1 J5 SUM fo r stage 4 IN ( 1 , 6 ) female pupa, day 1 SUM fo r stage 5 ADLT(l) adult male SUM fo r stage ADLT(2) adult female, day 1 6 As indicated in the equation above, on a given day 95 percent of the to ta l count in each element e lig ib le fo r graduation ( i . e . , the element represents a population which has been present in the given stage fo r at le a s t the number o f days indicated as the value of RATE fo r that stage) are accumulated into a sum and transferred into the day 1 position o f the succeeding stage (the remaining 5% stay in the element, to be subject to graduation conditions on the follow ing day). At the point o f graduation from fourth in s ta r, sex determination takes place and tra n sfe r is made into the male and female pupal stages (50% of the accumulated sum into each). Dynamics of the Egg Stage As there is a break in the system between the summer and spring egg components, a tra n s fe r of la s t year's viable eggs from the summer egg to the spring egg vector (both consisting of 24 elements represent­ ing the productive egg region o f the pool) is accomplished by the simu­ la tio n a t the beginning of the current season. Since the response of 62 the spring egg component is maturation to f i r s t in s ta r (shown graphically in Fig. IB ), egg hatch may occur when the program determines that the water temperature is > 5 °C and dissolved oxygen concentration is < 8 ppm. Other conditions fo r egg hatch are: FDO > PDO hatch = 0 FDO = PDO continue started hatch but do not in it ia t e new hatch FDO < PDO continue started hatch and i n it ia t e new hatch where FDO and PDO are the current and previous day's dissolved oxygen concentration, respectively. When hatching is determined to be possible, to ta l egg hatch is computed: NG [IE G G (I)] * [ 1 . 0 -0 .1 2 5 5 * FDO] * 0 .5 I=ITW According to the response equation h a lf o f the eggs are carried over to the next day (to simulate a delayed hatching e ffe c t) while the remainder e ith e r hatch as determined by [1 .0 - 0.1225 * FDO] * 0.5 or become non-viable. This is a lin e a r approximation expressing the percentage of eggs that hatch as a function of dissolved oxygen (Judson, 1960). 1 = 1 , 2 ,... IEGG( I ) is the number o f eggs in gradient I (where 24) and the sum to be transferred into the f i r s t in s ta r stage ( IN ( 1 )) is accumulated from a ll submerged egg containing gradients, i . e . , I takes values s ta rtin g from the index of the topmost submerged gradient, ITW, and continuing to the value o f NG (Index o f the deepest 63 egg containing gradient) by u n it increments. The value o f NG and the number o f submerged gradients on a given day are obtained from NG = Min (24, DM + 0 . 5 ) NWET = DEEP - XSDPTH + 0.5 where the addition of 0.5 allows a gradient one-half to completely flooded to be designated as submerged. Since egg gradients are assumed to be re s tric te d to the upper 24 one-inch le v e ls , and pool maximum depth (as input v a ria b le , DM) may range up to 36 inches, XSDPTH is a measure o f water depth below th is productive egg region. This quantity o f water is subtracted from the current pool depth (DEEP) with the resu ltan t integer number o f submerged egg gradients designated as NWET. Dynamics of the Adult Stage Adult dynamics as defined by the simulation are diagrammed in •.TSU fll---- Figure 1C and expressed MAL + MAL* (1 .0 - KILL (7 )) r FEM (1) = FEM (1)+FEM ( 8 ) FEM (D) = FEM ( D - l) * ( 1 . 0 - KILL (7 )) where D goes from eight to two by decrements o f one k FEM (1) = 0 where MAL and FEM are the male and female portions o f the population. The equations representing dynamics of the females ca lcu late the number 64 of individuals av a ilab le fo r advancement through the eight word array, advances the cohort to the next position in the array subject to m o rta lity , and sets the f i r s t position to zero in preparation fo r subsequent adult female input. Egg input response is computed by subroutine LAYIT fo r females positioned in the seventh position to simulate a seven-day gonotrophic cycle. The males simply remain in the f i r s t element position of the vector designated fo r that portion of the population. Both adult components are reduced by a d a ily m o rtality fa cto r KILL (7 ). The eggs la id by the females are designated as summer eggs and positioned in a vector (NEGG) of 24 elements (corresponding to one-inch levels along the flo o d lin e of the simulated woodland pool). Each day the eggs la id (to ta le d in IEGG) are dispersed evenly over the six gradients closest to the receding water level according to the following: t NEGG(I) = NEGG(I) + IEGG/ 6 i f condition 1 IEGG/number of dry gradients i f condition 2 i The conditions are 1. I f IBD > 2. I f 0 < IBD < 6 then I = IBD - 5, IBD 6 then I = 1, IBD where IBD equals the index o f the bottom dry pool gradient. MODELING THE PHYSICAL COMPONENTS Modeling the physical components of the system was accomplished a fte r consultation with individuals o f the National Weather Service at Michigan State U niversity and a review o f research on mosquito larval development in woodland pool ecosystems. The water depth component was updated d a ily u t iliz in g weather data from a data storage f i l e . The equation u tiliz e d in the simulation included an area facto r to represent the ra tio of microbasin area to surface area of the woodland pool that was studied during 1972 and 1973. The amount of water absorbed into the ground was a rough estimate pending results o f subsequent studies during th is same time period. The meterological features oriented as stim uli to the water temperature components are a ir temperature and soil temperature. These factors were considered to have measurable e ffe c t on the aquatic environment of the Aedes system. In research conducted by Haufe and Burgess (1956) i t was found that a ir temperature, wind c h ill and solar radiation were useful in a mathematical description of the relation ship between meteorological factors and the d a ily average temperature o f mosquito pools. region of subarctic Canada. The studies were conducted in a tundra Wind c h ill and solar radiation were not used in our model because the location o f the woodland pools in dense, forested areas and depressed microbasin areas was assumed to minimize the e ffe c t o f these meterological facto rs. 65 While solar rad iatio n was 66 i n i t i a l l y u tiliz e d as a behavioral feature to the water temperature component, the location of the aquatic microhabits allowed the parameter to be ignored in the la te r studies. F in a lly , the component fo r dissolved oxygen concentration was simulated with the following assumptions taken in to consideration. Since i n i t i a l l y the pool is f il l e d with runoff from m elting snow i t was f e l t that data contained in Welch (1952) on the oxygen content o f water saturated with a ir would provide a reasonable i n i t i a l value fo r the oxygen content o f the pool. The actual amount of oxygen d iffusing to the mosquito eggs (positioned in the bottom sediment) was modeled according to information in Ruttner (1963) expressing the rate of oxygen diffusing into the sediment as a function o f a constant A (eddy diffu sio n c o e ffic ie n t) and the range over which the concentration can vary. In the instance of a woodland pool ecosystem which in r e a lit y is an aquatic microbasin, the oxygen concentration is assumed to be uniform over depth. The actual mech­ anism fo r removal of oxygen from the pool was assumed to be microbial metabolism. This microbial uptake o f oxygen in our model was assumed to be a function o f temperature and an approximation o f data from Allen (1968) was used in the model to simulate oxygen removal from the pool. Dynamics of the Physical Components The dynamics of water depth is expressed and updated by the simulation in the following way: DEPTH = [DEPTH + PRECIP * RNF *AREAF - EVAP] [0.95 + 0.05 * RNF] 67 where RNF = 0.9 - 0.01*DAY. Runoff fra c tio n , RNF, into the pool was computed from a lin e a r approximation o f data obtained from the National Weather Service fo r average runoff proportion ( i . e . , 1 -proportion absorbed) in Michigan during the months o f March through June. I t was m u ltip lied by an area fa cto r (AREAF) of 10 to represent approximate r a tio o f microbasin area to surface area of the pool. A time lag of one day fo r 0.5 of the runoff from a day's p re c ip ita tio n has been added to the computer sim ulation. The term [ 0 .9 5 + 0.05*RNF] represented the calculatio n o f the proportion of water not absorbed into the ground and was our rough estim ate, pending results of subsequent studies. P recip itatio n (PRECIP) and evaporation (EVAP) are d a ily inputs into the simulation model and were obtained from data supplied by the National Weather Service a t Michigan State U niversity. The dynamics o f the component water temperature is expressed and updated by the program in subroutine WATERT. Once the routine determined that there was water present in the pool, the average water temperature fo r a 24-hour period was calculated based on the number of six inch water gradients present in a 36 inch maximum depth woodland pool (numbered from the bottom up). The previous day's average tem­ perature (TW) is set in a ll water gradients expressed as a six element vector, T l. Input variables are evaporation, p re c ip ita tio n , and water depth computed by a c a ll to subroutine WATER. The number (IX ) of six inch gradients f i l l e d with water was then calculated and the average temperature determined by the follow ing formulas where TA and TS equal 68 a ir and s o il temperature, resp ectively, and M goes from IX - 1 to decrements o f 1 2 by : WT = (T1( 1 ) + TA + TS*2)/4 I f IX < 1 T l(IX ) = TA+TW/2 T1(M) = [T1(M + 1 ) + T1(M - 1) ] /2 I f IX > 1 T l ( l ) = (T1( 2 ) + 2*TS)/3 The f i r s t condition states th at i f the number of water f i l l e d gradients is less or equal to one then the water temperature is equal to an average of the water gradient, a i r , and soil temperatures. The second condition states th at i f the number o f water f i l l e d gradients is greater than 1 , the top level is an average o f the a ir and water temperature while each subsequent gradient (with exception o f bottom gradient) is the average of the one above and below i t . The bottom gradient is the average of the soil temperature and the second gradient to the bottom. In both conditions 1 and 2 the soil temperature is weighted by a facto r of 2. F in a lly the average water temperature is weighted by volume of each gradient TW = AH/AV where AH equals the to ta l of volume X temperature fo r each gradient and AV equals the to ta l of volumes fo r the gradients. The component fo r the dissolved oxygen concentration of the woodland pool is calculated and updated in the follow ing manner. An 69 upper lim it in parts per m illio n o f oxygen in the pool is calculated as a function of water temperature PPM = MIN [1 4 .0 , 14.01 -0.242*WTEMP] where the term [14.01 - 0.242*WTEMP] was an approximation o f data in Welch (1948) on the oxygen content of water saturated w ith a ir at normal pressure. The d a ily dissolved oxygen le ve l is set at these values fo r saturation on days when p re c ip ita tio n occurs or on days when th is value produces a decrease in dissolved oxygen. The response variable FDO representing the level o f oxygen accessible to Aedes eggs in the bottom sediment is calculated as follows: SAT = 0.6*D0C UPT = [ 0 .2 1 +0.25*WTEMP] [12/1000] FDO = SAT - UPT The f i r s t equation represented the rate o f oxygen d iffu sio n in yg lite r" ^ hour"^ into the sediment of the woodland pool. The fa cto r 0.6 is the eddy diffu sio n c o e ffic ie n t of Ruttner (1963) w hile DOC is the current dissolved oxygen concentration in the water. The second equation represented an approximation of oxygen uptake in yg l i t e r hour"^ by microbial a c tiv ity (A lle n , 1968) m u ltip lie d by a fa cto r of 12 (to indicate our estimate o f the higher b acterial a c tiv it y o f a le n tic , depositional pool as opposed to A lle n 's work conducted in a lake system) and converted to parts per m illio n . The fin a l dissolved oxygen concentration av a ilab le to the eggs (FDO) is found by subtracting the value fo r UPT from SAT. 70 This completes the discussion of the computer program structure u tiliz e d in the description o f the dynamics o f the Aedes system. However, a description o f pool geometry was needed by the simulation to store the volume o f water contained in the pool and compute the number of one-inch gradients av a ilab le fo r egg input and hatch [Min (24, integer (max. w a te r ))]. The woodland pool microbasin is viewed as a segment from a large sphere th at is depressed from the surrounding land mass. The subroutine PONDIM computes the radius of this segment according to the following formula whereinput variables are surface area (SURFAR) and water depth (DPTH) of the pool at maximum flood le v e l: R = (SURFAR**2 + DPTH**2)/(2*DPTH)3 The pool is then divided into six-in ch gradients from the bottom up and the water volume stored in VOL(I) (where I = 1 , 2 , . . . 6) according to the follow ing segment volume formula where R is the radius of the sphere as calculated from the follow ing equation: SV = 1 * -rr * DEPTH * * 2 * (3 * R - DEPTH) PROGRAM STRUCTURE The overall program structure fo r the Aedes simulation ju s t discussed is diagrammed in Fig. 2. To provide f l e x i b i l i t y , a modular format was u tiliz e d th a t would accommodate additional variables and replacement or m odification of the o rig in a l equations used in the system. The mosquito's l i f e cycle is traced by the simulation in a d aily progression in it ia liz e d in the egg stage and developed through the immature stages u n til emergence as adults. The adult stage is then followed u n til expiration o f a ll adult mosquitoes with female egg deposition occurring during the adult cycle. When a ll eggs have hatched or become non-viable (due to environmental facto rs) and a ll of the other stages are empty, the w in terizin g phase established conditions for the next year's progression. The program also provides fo r addi­ tional years of simulation fo r long-term studies. The input variables fo r the computer simulation in two data f i l e s , WEATH and RUNDAT. The former is a weather f i l e containing data obtained from the National Weather Service a t Michigan State U n iversity. Currently a ir temper­ ature (max and m in), soil temperature, p re c ip ita tio n , evaporation, and runoff are being stored and updated on a d a ily basis fo r the months of March through August. These variables allow fo r the computation of values fo r the three physical components o f the Aedes system; water 71 72 temperature, water depth, and dissolved oxygen concentration o f the woodland pool. The RUNDAT f i l e contains data necessary to begin a program execution and contains information pertaining to egg lay and woodland pool conditions. Within each of the stages of development the inputs from the two f ile s control such processes as growth and maturation rate s , environmental m o rta lity , egg v ia b ili t y , and sex determination in the pupal stage. REFINEMENT OF THE MODEL'S BIOLOGICAL COMPONENTS f Immature and Adult M o rta litie s Used in Model The biological information on A. stimulans- f i t c h i i mosquitoes was implemented into the computer simulation model in order to re fin e the dynamics of the biological components o f the Aedes system. Immature m o rta lity rates of age in te rv a ls 7-28 are adequate estimations o f la rv al m o rta litie s with la te in s ta r m o rta lity approximately 0.50. However, i t was necessary to elim inate certain cohorts during the f i r s t age interval as s e ria l egg hatch (caused by additional flooding o f the woodland pool) introduced f i r s t instars into the la rv a l population resu ltin g in an underestimation of m o rta lity rates. This was especially true in 1973 when the 335 count fo r in s ta r I a t the second sampling t r i a l consisted e n tire ly o f newly hatched immatures and occurred a fte r the depth of the pool went from 18 inches to a maximum flood o f 36 inches. The elim in a­ tion of th is cohort from the l i f e ta b le resulted in more r e a lis t ic m o rta lity values fo r implementation into the model. These values are shown in the following table: SangHp 1 jnstar PoguUtjon 94 ^ 41 2 137 94 9 103 73 74 which gives a m o rta lity rate o f 0.248 (survival rate 0 .7 5 2 ). A s im ila r procedure was u tiliz e d fo r data collected in 1972 when 30 f i r s t instars were collected during the second sampling t r i a l and consisted o f nine newly hatched immatures. When these individuals were elim inated a m ortality rate of 0.230 (survival rate 0.770) was obtained. The revised to ta ls are shown in the following tab le: Samplinq T ria l In s ta r Popu lation I I I H I 1 26 35 2 21 14 Total 61 12 47 These values indicated th at e a rly development m o rta lity was approximately 0.25 fo r both years. Inspection of immature l i f e tables indicated that m o rta lity occurred p rim a rily during tra n s itio n from one in s ta r stage to the next. Although some d a ily m o rta lity is indicated, the data does not allow an adequate separation o f d a ily m ortality e ffe c ts . Survival rates used in the computer simulation are as follows: Daily Survival Transition Survival From in s ta r I to I I From in s ta r I I to I I I From in s ta r I I I to IV From in s ta r IV to pupae From pupae to adults 0.99 KILL( I ) = 0.01 0.75 0.75 0.50 0.75 0.50 and constitute an addition to subroutine GRADUAT. Although pupal to adult survival has no basis in the data, i t was estimated to be at least 0.50. This was indicated by the f ie ld studies th a t showed 75 highest m o rta lity occurred in la te r in s ta rs . In ad d itio n , empirical observations indicated th a t substantial numbers of adult mosquitoes fa ile d to completely emerge from th e ir pupal cases. Adult female m o rta lity rates used by the simulation were calculated using data in Table 7. In the model, day of emergence is considered to be the day when the larg est tra n s itio n from pupa to adult takes place. Complete survival of the adult female population is insured by the model fo r 36 days c fte r emergence. Thereafter rates fo r age in te rv a ls 36-85 in 1972 and 1973 were used to p lo t a composite survivorship curve (F ig . 3) th at established the female survival rates u tiliz e d in the simulation. The curve was plotted a fte r v is u a lly smoothing data fo r age in te rv a ls 36-64 to compensate fo r a divergent anomaly interpreted as due to external e ffe c ts such as predation, parasitism , or environmental stress. rate o f 0.965 (KILL(7) = 0 .0 3 5 ). This resulted in a d a ily survival A d a ily rate of 0.860 (KILL(7) =0.1 40 ) was established fo r age in te rv al 64 to 71 with the population decreased in the remaining two in te rv a ls by a fix ed number (1/14 o f the population at the s ta rt of the in te rv a l) u n til m o rta lity was 1.00. As empirical observations indicated th at adult males were present fo r approximately two weeks a fte r i n i t i a l emergence, they were retained in the model fo r 14 days at which time m o rta lity w ill be 1.00. Female Gonotrophic Cycle Refinement o f the computer model in lig h t o f acquired f ie ld data required us to review lite r a tu r e references in search o f a lte rn a tiv e 76 descriptions of individual biological components. In the case of female gonotrophic cycles fo r u n iv o ltin e , e a rly spring Aedes mosquitoes, Detinova (1968) lis te d results o f studies by Shlenova and Bey-Bienko (1962) showing the percentage o f gonotrophic cycles completed byj^k. communis (clo sely rela te d to A. stimulans- f i t c h i i ) during one season. Table 1 is derived from th at study and shows that approximately 0.3 of successive surviving portions of the female population w ill lay eggs on a weekly basis from the th ird week in June u n til the month of August. U tiliz a tio n o f th is information in the simulation resulted in a weekly egg input response expressed I EGG = (ADULTF * 0 .3 * 70.7) where IEGG = to ta l number o f la id eggs, ADULTF = to ta l number o f adult females, 0.3 = proportion of females undergoing an egg input response, and 70.7 = size of egg batch/female mosquito. For the purposes of the sim ulation, i t was not important whether th is egg input was from the f i r s t or succeeding gonotrophic cycles nor was i t lik e ly to have a detrimental e ffe c t to have a ll oviposition occur on one day of each week. For these reasons, the current version of the simulation w ill use a single variable location to represent the adult female component. Other factors re la tin g to size of egg batch w ill remain unchanged. The decision to incorporate a new procedure fo r simulating female gonotrophic cycles necessitated a revision of assumptions regarding geographic d is trib u tio n o f egg batches in the woodland pool. I f the egg lay procedure previously mentioned was retained, the low 77 level o f water a t the time o f egg lay would have resulted in most of the eggs being deposited in gradients 18-24. This was not supported by f ie ld surveys conducted in 1972 and 1973 which indicated substantial egg hatch when gradients 6-18 were flooded. Serial egg hatch was con­ firmed from f ie ld studies where the appearance o f newly hatched f i r s t in s ta r larvae was observed immediately a fte r inundation o f additional surface area of the microbasin. Also, the amount and location of the egg hatch observed at the study s ite confirmed published observations that JL stimulans eggs are generally la id in a regular d is trib u tio n across s o il gradients in a region below the maximum flood lin e to a depth o f 24 inches (H o rs fa ll, 1973). In 1973 no larvae were present in water samples taken when only the deepest portion of the microbasin was flooded (3 to 6 inch depth); so, fo r the model, an assumption was made th a t the female u tiliz e s an oviposition s ite th at is moist but not submerged (hence avoidance o f the pool bottom which remains flooded the longest) and a ll except the top six inches o f soil remain s u ffic ie n tly moist to be used fo r oviposition s ite s . Thus, in the simulation the female oviposits the eggs as a regular d is trib u tio n in the six to twenty-four inch gradients with the top six inches receiving 0.10 o f the to ta l egg lay. REFINEMENT OF PHYSICAL COMPONENTS In order to re fin e the computer sim ulation's description of those aquatic parameters used as key factors in immature mosquito dynamics, periodic measurements were conducted in 1973 on maximum and minimum water temperatures, water depth, and dissolved oxygen concentration. These measurements were taken during a 14 week period beginning March 1 and continuing through June 15 a t the woodland pool study area. In a d d itio n , d a ily measurements fo r meterological features oriented as stim uli to the physical components of the Aedes system (maxmin a ir temp., soil temp„ p re c ip ita tio n , and evaporation) were obtained at the aquatic m icrohabitat during the same time period. The data was placed on a storage f i l e fo r subsequent s ta tis tic a l analysis and use in the refinement o f the model. Water Temperature Analysis of meterological data showed that maximum water temperature had more influence on dissolved oxygen concentration than minimum water temperature (Table 2 ). I t is also known from Haufe's research (1957) th a t immature stages of a related species have a tem­ perature preferendum in a woodland pool ecosystem and w ill seek optimum temperature le v e ls . F in a lly when the temperature averaging method was used to pred ict water temperature (version 2 of the Aedes sim ulation), 78 79 larval development lagged behind the ra te observed in f ie ld studies with the re s u lt th at the immature population died before any of the la te in s ta r stages were reached. I t was due to these three reasons that maximum water temperature was used in place o f the temperature averaging method. A m u ltip le regression analysis was conducted on the f ie ld data to determine the best relation ship between maximum water temperature and a corresponding set of independent variables with the following re s u lt: MAXW = [-2.5215 + (0.8812*MAXAIR) - (0.0228*MINAIR) - (0 .0897*S 0IL T )+ (0.2710*YMAXW)] where MAXW = maximum water temperature, MAXAIR = maximum a ir temper­ ature, MINAIR = minimum a ir temperature, SOILT = soil temperature, and YMAXW = the previous day's maximum water temperature. The proportion of v a ria tio n explained by these c o e ffic ie n ts is 0.9520 with MAXAIR contributing almost 0.50 to the overall c o rrelatio n c o e ffic ie n t. Dissolved Oxygen Concentration Dissolved oxygen concentration and various combinations of measured temperatures were analyzed by step-wise m u ltip le regression to choose the best model in an attempt to re fin e the computer simu­ la tio n . A comparison between measured and simulated oxygen levels indicated th at the simulation was producing values in excess of actual concentrations obtained in the f ie ld . The most accurate prediction equation which emerged from the data analysis was as follow s: 80 FDO = [3.1374 - ( 0 . 1798*MAXW) + (0.2630*DOC)] where FDO = dissolved oxygen concentration av ailab le to eggs fo r hatching stimulus, MAXW = maximum water temperature, and DOC = d is ­ solved oxygen concentration of the pool. The c o e ffic ie n ts o f this equation indicate th a t a smaller eddy c o e ffic ie n t should be used (0 .3 instead of 0 .6 u tiliz e d in the simulation model) and the value fo r increased uptake of dissolved oxygen by bacterial metabolism. These changes, in addition to the replacement o f average water tem­ perature with maximum water temperature, produced values in the model that were more compatible with f ie ld observations. The proportion of v a ria tio n explained by these c o e ffic ie n ts is 0.4312. There is reason to believe th a t change in oxygen levels rather than absolute levels is the stim ulating fa cto r in egg hatch (Judson, 1962). The model now uses change in oxygen level to stim ulate hatch but u tiliz e s an equation derived from the previously mentioned author's work on s ta tic oxygen levels and th e ir e ffe c t on tra n s itio n from the egg stage and survival to in s ta r I . This is , no doubt, resu ltin g in a reduced hatch in the computer simulation but in s u ffic ie n t data con­ cerning the mechanics of th is process negates any change in the model at th is time. Water Depth A c o rre la tio n of 0.8391 was obtained fo r the simulated and actual f ie ld data which indicated a reasonable model fo r seasonal 81 inundation o f the microbasin area. Correspondence between d a ily values varied during the study period with simulated depth increasing and decreasing a t a fa s te r rate than observed f ie ld conditions. This suggested th a t real pool depth depends, to a large degree, on water table levels during spring flood conditions. This is in agreement with H o rs fa ll's (1963) contention th a t the level o f the water table is an important fa c to r in the flooding o f a microbasin area and, in add itio n , the a b ilit y to retain additional input o f water as a re s u lt of snow melt and/or p re c ip ita tio n . No data was ava ilab le concerning flu c tu a tio n of water tab le depth at the study s ite so substantial refinement o f th is component was not undertaken. 82 Footnotes Journal A rtic le No. 6989, Michigan A g ricultural Experiment Station. This research was supported by a fellowship grant from the Michigan State Department o f Natural Resources. 2VARIABLE NAME used throughout the paper are the actual variable names u tiliz e d in the computer program. 3Since we are using FORTRAN language fo r the equations contained in th is paper, (VARIABLE NAME)** X represents a given quantity raised to the Xth power. 83 Table 1. Age structure o f a population o f u n ivo ltin e Aedes mosquitoes3 Proportion o f Population Completing Indicated Number o f Gonotrophic Cycles Month June July August 0 1 3 0.7 0.3 0.3 4 0.4 0.6 0.3 1 0.1 0.9 0.3 Week 2 3 4 Proportion of Population Laying Eggs 0.4 2 0.7 0.3 3 0.6 0.3 0.1 4 0.4 0.2 0.3 0.1 0.6 1 0.3 0.3 0.2 0.2 0.2 0.2 aModified from Shlenova and Bey-Bienko, 1962. 84 Table 2. C orrelation m atrix fo r max-min a ir and water temperatures and dissolved oxygen concentration in a woodland pool ecosystem a t Yankee Springs recreational area, 1973 Dissolved Oxygen L" Max Water Temp Min Water Temp Max water temp -0.5923 Min water temp -0.3792 0.7526 Max a ir temp -0.5090 0.8775 0.7107 Min a ir temp -0.4414 0.7260 0.8542 Max A ir Temp 0,8285 85 Fig. 1. Flowchart of Aedes population dynamics as expressed by the computer simulation model. stages. B. A. Population dynamics of immature Location o f key pool dimensions and spring eggs and egg hatch response. C. Dynamics of ad u lt population, egg input response, and location o f summer eggs. Subroutine DEVLOP moves to ta ls in a ll vectors (except egg stage) down from position 1. Subroutines HATCH, GRADUAT, and LAYIT tra n s fe r cohort sums as shown by arrows. MAXIMUM DEPTH MALE ADULT VECTOR SUMMER EGG VECTOR SPRING EGG VECTOR I= k EGG INPUTRESPONSE FEMALE ADULT VECTOR •CZH-j IBD 24 -EGG HATCH RESPONSE DEEP NG XSDPTH PUPAL VECTORS MALE INSTAR I VECTOR INSTAR 2 VECTOR INSTAR 3 VECTOR INSTAR 4 VECTOR ♦ iC Z] •CZD RATEI4) RATE(2) RATE ( I ) 20 FEMALE RATEI5) RATEI3) 20] 20 20] A— Figure 1 RATE(6) 20] 87 Fig. 2. Flowchart of the program structure of the Aedes computer simulation model. rnmmmmmmrnm INITIALIZATION ' P 0 « . DIMENSIONS EGG INPUT LOCATION WINTER INPUT PREFLOOD CONDITIONS INtTALIZE PARAMETERS, NO ADULT EMERGENCE END NO EGGS ENO YES YES ADULTS NO YES WEATH FILE GENERATE/INPUT NO CONTROL FACTORS YES COMPUTE COMPUTE MORTALITY FOR EACH STAGE EGG INPUT RESPONSE COMPUTE GROWTH RATE FOR EACH STAGE ENO OF YEAR SUMMARY DATA OUTPUT ADVANCE COHORT IN EACH GRADUATE COHORT TO NEXT STAGE ® Figure 2 CYCLE END OF SEASON 89 Fig. 3. Survivorship curves for A. stimulans- f i t c h i i adults a t Yankee Springs recreational area, 1972-1973 d iv is io n s ). (sem i-log, 2 cycles x 10 500 No. Survivors (Log S c a le) 1000 \ SURVIVORSHIP CURVE FOR 1972 \ SURVIVORSHIP CURVE FOR 1973 SURVIVORSHIP CURVE FOR USE IN SIMULATION I 00 50 _L 0 -3 6 I 43 l 50 I 57 i 64 Age Interval (Days) Figure 3 71 78 SUMMARY AND CONCLUSION Mosquito b itin g a c tiv ity was studied in Michigan parks from 1971 to 1973. The results indicated th a t woodland Aedes mosquitoes were the pest species in the m ajority of state parks. Adult collection s a t North Higgins Lake Park and Yankee Springs recreational area during 1972 and 1973 re v e a le d ^ , communis- tric h u ris and A_. stimulans- f i t c h i i species complexes, resp ectively, were the major mosquito pests. Popu­ latio n s of the former species complex were r e la tiv e ly short lived with highest b itin g a c tiv ity occurring la te May. Populations of A. stimulans- f i t c h i i existed u n til la te August with highest b itin g a c tiv ity occurring la te June and July. Preliminary observations on e ffe c t on human recrea­ tio n a t the two parks showed planned recreational a c tiv itie s were cancelled during periods o f high mosquito b itin g a c tiv ity . Studies were also conducted during 1972 and 1973 a t Yankee Springs on population dynamics ofjA . stimulans- f it c h i i mosquitoes. The results of f ie ld studies indicated extreme d a ily v a r ia b ilit y in immature mosquito dynamics as well as weather conditions from March 1 to A pril 15. Examples included s e ria l egg hatch over a 14 day period resu ltin g in the addition of f i r s t in s ta r cohorts and an immature popu­ la tio n consisting of a ll four in s ta r stages during the month o f A p ril. Biological and physical parameters were much more stable in la te A pril and e a rly May. The water depth in the pool was stead ily receding and 91 92 a ll cohorts were in the la te in s ta r stages of development and uniform in d is trib u tio n and composition. The former phenomenon was due to the progressively decreasing pool depth which constricted the larva l population into a smaller geographical area. The more uniform in s ta r composition previous to adult emergence was due to the warmer water temperature which accelerated growth rates o f a ll immature cohorts. Seasonal inundation of the microbasin area was consistent fo r the two years with the highest water levels measured in 1973. This was a fa cto r in determining immature population densities as the higher population to ta ls in 1973 were due to the flooding o f additional egg containing soil gradients of the pool. Overall m o rta lity was approximately 0.75 fo r both years with the highest rates observed in the la te in s ta r stages. As the rate o f larva l development and biomass increase was rapid in a temporary aquatic ecosystem, n u trie n t re s tric tio n s placed on immature mosquito populations would explain the substantial a t t r it io n during la te r developmental stages. Adult female populations were comparable fo r the two year period with s lig h tly higher to ta ls in 1973. Peak population levels were maintained fo r approximately one month a fte r emergence with substantial m o rta lity occurring during July and August. A p lo t of survivorship curves fo r 1972 and 1973 showed th at the A. stimulansf it c h i i female population had minimum m o rta lity fo r most of the adult l i f e span. High losses occurred p rim arily among the older females which indicated m o rta lity rates were age dependent. The significance of this data is th at the a b ilit y o f th is mosquito to e x is t a t high 93 population levels fo r so long made i t a potential disease vector as well as the insect pest responsible fo r continual disruption of recreational a c tiv itie s throughout the summer months. The results o f f ie ld studies conducted a t Yankee Springs recreational area on populations of_A. s tim u la n s -fitc h ii mosquitoes were used in v a lid a tio n of the computer simulation. This was the process o f gaining confidence th a t the model represented its real counterpart in the natural ecosystem. Once the model was refined with data from the study, deviation between the simulated and observed larval responses was small enough to re s u lt in adequate correspondence o f the model's intern al structure to th at of the real system. Simula­ tion values fo r the biological components representing the immature stages showed the d a ily v a r ia b ilit y from March 1 to A pril 15 th a t was observed in the actual f ie ld studies. included: Examples o f the model's responses Twofold increase in water depth during 48 hours; s e ria l egg hatch over a 14 day period resu ltin g in the addition of fiv e f i r s t in s ta r cohorts; and an immature composition o f a ll four in star stages in e a rly A p ril. However, the modeling approach indicated deficiencies in research and/or lack of information on the actual dynamics of certain biological components o f the Aedes ecosystem. This was especially true of egg-laying behavior and la rv a l and egg m o rta lity . The model's approach to these behavioral features was purely speculative and im plications fo r the design o f experiments to obtain improved estimates are obvious. The same can be said fo r simulation values representing physical features of the woodland pool. The low proportion o f v a ria tio n 94 explained by the model fo r dissolved oxygen concentration indicated the dynamics o f th is aquatic feature was not well understood and resulted in a reduced egg hatch in the computer sim ulation. Values fo r oxygen levels were probably being underestimated but in s u ffic ie n t data precluded an improved parameter estim ate. Although there was a high c o rre la tio n between simulated and actual values fo r water depth, d a ily correspondence between these values varied with simulated depth increasing and decreasing at a fa s te r ra te . The model did not take into consideration the level o f the water ta b le , an important facto r in the flooding of a microbasin area. Experiments conducted to under­ stand flu c tu a tio n o f water table le ve ls during spring flooding would re s u lt in substantial refinement o f th is component. As a re s u lt pre­ dictions on immature la rv a l populations in a given year could be made with greater precision. The strong point o f th is modeling process was th a t fo r the f i r s t time an attempt had been made to look at the l i f e cycle and physical parameters o f woodland Aedes mosquitoes through a systems science approach. Research in the past has taken a haphazard approach with numerous research projects concerned with various aspects of the l i f e cycle but no attempt was made to show how the results related to the e n tire l i f e cycle. The Aedes model is so constructed th at i f an assumption concerning a component proved in v a lid , th a t portion of the model can be replaced by the correct functions. This modular approach insures th a t the model w ill remain a useful working to o l. the system One use of model in, the future w ill be in the area o f pest management. 95 Prelim inary analysis o f the computer simulation has shown that immature mosquito larvae are quite sensitive to a lte ra tio n o f physical parameters o f the aquatic ecosystem. What makes s e n s itiv ity analysis so in te re s t­ ing is th at i t considers the Aedes system as a whole and very often a change in one system parameter produced an unexpected change in another. As a r e s u lt, s e n s itiv ity analysis can provide a framework fo r pest management decisions. Another future consideration is the use o f the simulation in the study o f v ir a l m u ltip lic a tio n in the mosquito. L ittle information is av a ilab le concerning the mechanism o f v ir a l buildup in the invertebrate vector. A model could be constructed u t iliz in g current research data and coupled to the woodland Aedes model to simulate v ira l a c tiv ity in the mosquito. As results o f subsequent research become a v a ila b le , more v a lid functions can be substituted. LITERATURE CITED LITERATURE CITED A llen , H.L. 1968. Acetate in fresh water: natural substrate concentrations determined by d ilu tio n bioassay. Ecoloqy 49:346-349. Anderson, J.F. and IJ.R. H o rs fa ll. 1969. Thermal stress and anomalous development of mosquitoes (D iptera: Culicidae) I . E ffe c t of constant temperature on dimorphism o f adults of Aedes stimulans. J. Expt. Zool. 154:67-89. Barr, A.R. 1958. The mosquitoes of Minnesota. Univ. o f Minnesota Agr. Exp. Sta. Tech. B u ll. 228. 154 pp. Brylinsky, M. 1972. Steady-state s e n s itiv ity analysis of energy flow in a marine ecosystem, Part I I , pp. 81-101. l£ B.C. Patten (e d .), Systems Analysis and Simulation in Ecology. Academic Press, New York. Vol. 2, 592 pp. Carpenter, S.J. and W.J. LaCasse. 1955. Mosquitoes of North America. Univ. of C a lifo rn ia Press, Berkeley, 360 pp. Caswell, H ., H.E. Loewig, J.A. Resh and Q.E. Fuss. 1972. An in tr o ­ duction to systems science fo r ecologists, Part I , pp. 4-76. B.C. Patten (e d .), Systems Analysis and Simulation in Ecology. Academic Press, New York. Vol. 2, 592pp. Dayton, J.R. 1971. Bionomics of the mermithid parasite of Aedes stimulans (Walker) (D iptera: C u lic id ae ). M.S. thesis. Mich. State U n iv ., 98 pp. D eF oliart, G .R., M.R. Rao and C.D. Morris. 1967. Seasonal succession of bloodsucking Diptera in Wisconsin during 1965. J. Med. Entomol. 4:363-73. Detinova, T.S. 1968. Age structure of insect population o f medical importance. Ann Rev. Entomol. 13:427-450. Gorton, R.J. 1973. An epidemiological study o f mosquito-borne C a lifo rn ia encephalitis in a southern Michigan subdivision. M.S. thesis. Mich. State U n iv., 75 pp. Hagan, N.K.B. 1966. Mermithid nematode parasitism of Aedes stimulans (Walker) (Diptera: C ulicidae) from Ingham County, Michigan. M.S. thesis. Mich. State U n iv., 65 pp. 96 97 Haufe, W.O. 1957. Physical environment and behavior of immature stages of Aedes communis (Peg.) (D iptera: Culicidae) in Subarctic Canada. Canadian Ent. 89:120-139. H o rs fa ll, W.R. 1963. Eggs o f floodwater mosquitoes (Diptera: Culicidae) IX. Local d is trib u tio n . Ann. Entomol. Soc. Am. 56:426-441. H o rs fa ll, W.R., P.T.M. Lum and L.M. Henderson. 1958. Eggs of floodwater mosquitoes (D iptera: Culicidae) V. E ffect of oxygen on hatching o f in ta c t eggs. Ann. Entomol. Soc. Am. 51:209-213. H o rs fa ll, W.R. and H.W. Fowler. 1961. Eggs of floodwater mosquitoes. V III. E ffect o f s eria l temperatures on conditioning of eggs o f Aedes stimulans (W alker). Ann. Entomol. Soc. Am. 56:644-66. H o rs fa ll, W.R. and M. Trpis. 1967. Eggs of floodwater mosquitoes. X. Conditioning and hatching o f w interized eggs o f Aedes s tic tic u s (D iptera: C u licid ae). Ann. Entomol. Soc. Am. 60:1021-1025. Irw in , W.H. 1942. The role of certain northern Michigan bog mats in mosquito produc'tion. Ecology 23:466-477. James, H.C. 1961. Some predators of Aedes stimulans (Walker) and A. trichurus (Dyar) in woodland ponds. Can. J. Zool. 39:533-40. Judson, C.L. 1960. The physiology o f hatching of aedine mosquito eggs: hatching stimulus. Ann. Entomol. Soc. Am. 53:688-691. McDaniel, I.N . 1958. Bionomics o f the brown woods mosquito, Aedes stimulans (Walker) (D iptera: C u licid a e). Ph.D. thesis. Univ. of 111. , 112 pp. Patten, B.C. (e d .). 1972. Systems Analysis and Simulation in Ecology. Academic Press, New York. Vol. 2, 592 pp. Pederson, C.E. 1947. The d is trib u tio n o f Michigan mosquitoes. M.S. thesis. Mich. State U n iv ., 83 pp. Ruttner, F. 1963. Fundamentals of limnology. Toronto Press. 295 pp. 3rd ed. Univ. of 98 Shlenova, M.F. and 1.6. Bey-Bienko. 1962. Age composition of mass species populations o f mosquitoes genus Aedes (According to observations made in Byelorussia). Probl. Gen. Zool. Med. P a ra zito l. 589-605. Trpis, M. and W.R. H o rs fa ll. 1969. Development of Aedes s tic tic u s (Meigen) in re la tio n to temperature, d ie t, density and depth. Ann. Zool. Fennici. 6:156-160.