Habitat Selection Strategies of Mosquitoes Inhabiting Citrus Irrigation Furrows

JuNn, 1985 J. Ar"r.Mosg. CoNrnor Assoc. 169

HABITAT SELECTION STRATEGIES OF MOSQUITOES INHABITING CITRUS IRRIGATION FURROWS

G. ALAN CURTISI

ABSTRACT, The mosquitoes Aedcs uexans, Psorophora columbiae and Psorophora howardii have recently extended their habitat distiibution into citrus irrigation furrows in coastal southeastern Florida. Oviposition site selection was examined by correlating species abundances with water depths due to rainfall or flood trigation. Psorophora eolumbiae and Ps.howardii oviposited low in furrows, shared similar distributional relationship"s to water depths and were hatched by rainfall or irrigation. Aedes texans oviposted higher in the citrus furrow, showed a different relationship to water depth and were hatched only by flood irrigation.

INTRODUCTION tics for colonization and interspecies relation- ships used by mosguitoes inhabiting citrus irri- Presence of mosquito larvae in any habit is gation furrows. determined entirely by maternal choice. Habitat selection is among the most important behavioral responses that *re female mosquito MATERIALS AND METHODS makes. since the effect of this choice dictates the from two fate of her offspring. Successful choices by Data presented here were collected groves located ca. pioneering females may result in establishment similar 16.2 ha mature citrus River County,FL(27.5 of species in a newly exploitable environment 16.l km apart in Indian by the (Curtis and Frank l98l) or perpetuate the spe- N, 81.5 W). Each grove was cultivated crown flood cies in an established habitat. Ifselected habitats same techniques, which included is given contain inadequate food, abundant predators irrigation. A general grove description by Curtis and Frank (1981). or insufficient water then few or weak individu- were collected from als may result (Krebs 1972). Biotic and abiotic data period (1978-82)' Successful mosquito establishment in man- the 2 groves over a 5-year made habitats is of particular interest. Species Following rainfall or flood irrigation, immature sites that are able to select and propitiously adapt to mosquitoes were collected from 9 selected by a a man-made habitat have the possibility of within the citrus grove, determined An open ended 0.25 exploiting a new environment, with few or no stratified random method. bottom substrate other mosquito species present. One of the m2 box was forced into the pupae removed for commonest man-made mosquito habitats is that and all larvae and were (Curtis and Frank of land irrigated for agricultural purposes counting as earlier described as long as (Al-Azawi and Chew 1959, Boyd 1941, Briscoe l98l). Collections were made daily 1952, Edmunds 1958, Gunstream and Chew the aquatic habitat persisted. During each sam- 1964, 1967; Harden, et al. 1967,Hill et al. 1977, pling occasion when sampling was conducted, with Portman 1954. Reeves and Hammon 1962, water depth was recorded in association from Shemanchuk 1959 and Surtees 1970a, 1970b). mosquito species abundance, whether However, the mechanisms for mosquito estab- rainfall or flood irrigation. mos- lishment in man-made habitats are not well Soil samples for locational detection of known. quito eggs were collected as described by Curtis (1981). Citrus groves in coastal southeastern Florida and Frank Samples were analyzed from 3 have been intensively cultivated since the early specific vertical stratifications of the citrus irrigation (A) (bot- 1900s, but the numbers of associated mos- furrow, these being: lower (B) quitoes have increased relatively recently (Cur- tom of furrow to 0.3 m vertical elevation), - (C) tis and Frank l98l). Much of this increased middle (0.3 0.6 m elevation) and upper - abundance is in response to the advent and section (0.6 0.9 m elevation) of the furrow widespread use of crown flood irrigation. In a (Fig. l). From each location 30 random samples previous study employing egg survey tech- were processed and mosquito eggs identified. niques, it was demonstrated that Aedes oexans The ability of rainfall or flooding to fill and oviposited in a specific region of the citrus fur- persist in the furrow was evaluated by measur- row (Curtis and Frank l98l). ing water depth at randomly selected sites The present study describes some of the tac- throughout the grove following rainfall or flooding. Sampling was continued as long as measurable water persisted in the habitat. For I Indian River Mosquito Control District, P.O. Box the purpose of correlating habitat persistence 670, Vero Beach, Florida, 32961. with a given rainfall, only the occasions when no 170 J. Ar"r. Mosq. CoNrnor Assoc. Vor. I, No. 2

intervening rainfall occurred were analyzed. Re- 12 gressionanalysis on data to produce figures 2-5 was accomplished using a best fit program for linear and nonlinear data using least squares techniques(Daniel and Wood 1980). aA

RESULTS t- The standard configuration for a citrus furrow in coastalFlorida resemblesa parabola with a height of 0.9 m and width of 7.6 m. Figure I depicts the amount of rainfall in 24 hr neces- sary to fill the furrow to various depths and the o216aroiz (CU) percentage of the toral iltnF 4 depth filled by rainfall. Fig. By design,crowrr-flood 2. Regression of rainfall and number of days irrigation fills the fur- that free water persisted row to l00Vo in the furrow following spc- capacity, a value that would re- cific rainfall amounts. Regression quire described b-v ihe over 43 cm of rainfall in 24 hr. equation, y=abr (R2=0.862, n= 125).

O.9 n

\ too X (44.7 cn) ,l .6 77 Z (37 .9 cn) \ ./

57 Z @A.7 cn) e m \, ./

\ ze z (n.5 cd 1/ (A) \- 6 % (7.5 cn) .A % (2.o cn)

Fig' ' Citrus furrow profile ^l - showing physical dimensions and relationship between furrow configuration and rainfall required to fill it to various depths. Vertical sections A, B and Ciesignate areas used fo-r special soil samples.

The relationship between amount of rainfall through May for Ps. colurnbiae and Ps. howardii within 24 hr and persistence of water in the when eggs of these mosquitoes are in a quies- furrow is described by a power curve equation cent period. (y : abfl, R2 : 0.862, P<0.001) (Fig. 2). In Analysis of Ae. uexans eggs collected from soil contrast, standing water persists on average l7 samples-gave means of 0.3, 2.0 and 43.3 eggs/ days following flood irrigation. sample from sections A, B and C respectively. During the study period, llg,5l8 Aedcsaexarc The non-parametric rank comparison (Dunn's ( Meigen), 43,37 6 P sm ophora co lurnbiae (Dyar and Multiple Comparison Technique (Dunn 1964)), Knab) and 1,394 Ps. howardii Coquillett were concluded that section C, the uppermost re- collected in 1256 water samples. The abun- gion, contained significantly greater Ae. uexans dance of each species showed a significant cor- egg numbers than the other 2 sections relation with water depth, whether from rain- (P<0.05). Psorophora cohtrnbine egg collections fall or irrigation (figures 3-5). These species produced means of 5.2, l.0 and 0.I eggs/sample represented 97.3Vo of all mosquitoes collected for sections A, B and C respectively. Statistical during the course of study. Figures 3-5 repre- analysis confirmed that section A, the lowest sent the mean number/sample from the 9 region, contained significantly greater numbers collections/day following a specific rainfall or of Ps.columbiae eggs than B or C (P<0.05). The flood irrigation. mean number of eggs/sample for Ps. howardii Excluded from the data are the occasions of was 1.93 for A, 0.33 for B and 0.10 for C with A rainfall or flooding during late December being significantly greater than B or C(P<0.0b). Junr, 1985 J. Aru. Mosq. Coxrnor Assoc. t7l

2AO

260

210

700 f, zoo 5 I reo I eoo 160 \ \ 5oo E reo j, eo > 400 AAI AA 2 too z z Z loo

t-- I 60 200 10 too 20

R DEPTH (CM)

Fig. 3. Regression of water depth from either rain- Fig. 5. Regres,,"" "r;;*.::;; *"- either rainfall or flood irrigation and mean/sample of Psorophma fall or flood irrigation and mean/sampleof Aedesvex- columhiat. Regression described by cubic equation, azs is described by the equation, y:0.95eo.osx _23.6+ _0.lgxr+0.001 ya y: I l.9x 1pz:0.96, (R,:0.79, n:1,256). n:851). Samples during egg quiescent period have been excluded. immature Ae. vexans remained low up to a l5 depth of 50-60 cm whereupon a steep increase in Ae, aexans numbers was observed. !3 12 5 rt DISCUSSION > to The successful exploitation of citrus groves a!

t by mosquitoes is a clear example of a man-made ?e environment being colonized by indigenous t- >4 species. In the case of Ae, uexans, an uncommon resident mosquito became abundant because it 2 was able to take advantage of the recently I o created habitat (Curtis and Frank l98l). o 20 "o !o Psorophora columbiae and Ps. houardii have "orf o."r" ,"u, selected other regions of furrows to fit their specialized developmental needs. These mi- Fig. 4. Regressionof water depth from either rain- crohabitats seem to mimic their traditional nat- fall or flood irrigatioir and mean/sample of Psorophma howardii. Regression described by cubic equation, ural habitat. y=0.08+0.4x-0.01 x2+2.4E-bx3 (Rr:0.92. Due to the ephemeral nature of the aquatic n=851). Samplesduring egg quiescentperiod have citrus furrow habitat, desiccation is the single been excluded. largest contributor to mosquito mortality. Fur- row construction is such that drainage is facilitated by a slight gradient along its length. The curvilinear relationship between water Therefore, larvae residing at upper elevations depth, whether from rainfall or flooding and along the furrow's length may face desiccation Ps. colurnbiae, (Fig. 3) indicates that there is a sooner than those in the lower end of the fur- relatively steady increase in mosquito numbers row. Desiccation was the principal cause of with .increasing water depth to a level of ap- mortality in artificially flooded ponds as re- proximately 35 cm, at which point the curve ported by Pritchard and Scholefield (1983). becomes asymptotic, as described by the cubic Aedesuexaru, Ps. columbiae and Ps. howardii have regression of water depth and Ps. colurnbiae minimized this risk by selecting differing -23.0+ -0. (y: t l.9x 19x2+0.001xsRz:0.86). oviposition sites that potentially insure larvae A third degree polynomial expression demon- sufficient development time. Mclachlan and strates a similar relationship for Ps. howardii Cantrell (1980) reported dipteran developmenr -0.01 (y:0.08+0.4x x2 +2.4F'-bx,3 Rr:0.92) time and habitat transience were positively cor- wiqlr the exception that the asymprote begins at related in rock pools. a slightly shallower depth than for Ps. colirnbiae The assumption that larval collections taken (Fig. a). A different association exists between in association with varying water depths is an water depth and Ae. uexans abundance indicator of oviposition proved accurate. Albeit (y:0.95er, R,:0.79) (Fig. 5). Collections of the method is indirect in comparison to egg t72 J. Attl. Mosq. CoNrnol Assoc. Vou l, No.2 sampling, the statistical results (figures 3-5) and not abundance. The comparison of larvae allow little probability in making a type II error collected in relation to water depth via the in acceptanceof the wrong hypothesis. The re- Kolmogorov-Smirnov test, demonstrated no gression equations for the three mosquito spe- significant difference in distributions between cies all provided an excellent fit to the individ- thesetwo mosquitoes(P >0.25). ual data sets.The profiles of water depth and The distribution of Ae. t)exons,due to oviposi- mosquito abundance appear to be a good indi- tional selection, contrasts sharply with that of cator of oviposition selection. Eggs collected the Pswophoraspecies. This flood plain mos- from the soil samplescorroborate the locations quito (Horsfall et al. 1975)shows a pronounced predicted by the regressionanalysis. Eggs ofAe. exponential relationship to increasing water oexanswere collected in the upper 30 cm of the depths (Fig. 5). Relatively few individuals were furrows as also witnessed from the 1,256 collectedafter most rainfalls. whereasAe, vexans larvae-water samples.Following rainfall, larvae were numerous after flood irrigation or the of Ps. colurnbiaeand Ps. howarilii were most rare rainfalls in excessof 29 cm in 24 hr. abundant in the lower 30 cm range. Eggs of Aedesuexans exhibits the longest development these two species were similarly abundant in time of these 3 floodwater species,with a mean this sameregion. This distribution of larvae and of 8.0 daysand range of 5.0- 17.0days in citrus eggs, as detected by hatching response to vary- groves (Curtis, unpublished data). Only rainfall ing water depths and soil samples, illustrates in excessof 10.2cm would assurethis mosquito female oviposition site selection within the cit- ample time for development (Fig. 2). Although rus furrow habitat. 10.2 cm of rainfall might insure ample habitat Both Ps. columbineand Ps, howardiioviposit in duration, Fig. I illustrates that this amount of the lower regions of the citrus furrows. Normal rainfall is not adequate to fill the furrow to Ae. rainfall or flood irrigation will immerse eggsof uexanspreferred zone (C). Here, other oviposi- these species. After eclosing, they progress tional cues, such as shading and leaf litter are rapidly to adults, with mean development times more influential. (and ranges)in citrus groves of 5.5(4.08-8.0) Comparison of Ae. aexansdensity at different daysand 4.5(3.5-8.0) daysforPs. columbi,aeand water depths with that of Ps.columbiae and Ps. Ps. howardi.irespectively (Curtis, unpublished howardiidemonsrated significant differences in data). SincePs. howardiiis an obligate carnivore distributions (Kolmogorov-Smirnov test after the first instar, it must rely on the con- P<0.001) among both Psorophoraspecies and comitant hatching of prey mosquitoes or face Aedesaexaw. Previous results indicate that Ae. certain starvation or desiccationif temperature uexansselects oviposition sites in the leaf litter, or insufficient prey extend development time high on the citrus furrow under the canopy of beyond the habitat's permanence. the citrus trees where egg hatch is only stimu- From figures 3 and 4 it is seen that Ps.howar- lated by crown flood irrigation or excessive dii and Ps.colurnbhtc peak in abundance at water rainfall (Curtis and Frank l98l). This permits a depths of30 and 35 cm respectively,equivalent long development time from egg hatch. Water to 14.8cm and 17.3cm of rainfall. Flooding by from flood irrigation persistsan average of l7 irrigation produces water depths between 60-90 days, occasionallyin an excessof 20 days. This cm, but does not increasenumbers of Ps.howar- period is sufficient for successful metamor- dii or Ps.colu,mbiae. If thesetwo mosquitoesused phosis of Ae. aexanseven when lower water the entire available citrus furrow for oviposi- temperture slows advancement. tion, there would be a continuous increasein Aedesvexans may be collected at any time of numbers with increasing water depths. the year when flood irrigation occurs, but eggs Psorophmacolumbiae and Ps. houardii are using of Ps.howard;ii and Ps. columbiaedo not respond only about 20Voof the available furrow surface to ordinary hatching stimuli between late De- area, specifically, the lower zone, that most cember through early May. This quiescent pe- likely to be inundated with normal rainfall. riod is advantageous, since egg hatch in re- Here, due to the parabolic shapeof a furrow, sponseto normal winter rainfall would present mosquitoes encounter the moist, often satu- eclosing larvae with an environment of cool rated, soil. Such a substrate is highly attractive water temperatures. The resulting develop- to ovipositing Ps. columbiae(Olson and Meek ment times would exceed the aquatic habitat's 1977,Al-Azawiahd Chew 1959).I-ower regionsof permanence, and result in death due to desic- the furrow are most similar to thoseof pasturesand cation. Lower water temperatures would also low lying environments commonly used by Ps. slow the predatory rate of Ps.howardii even if colurnbinz and Ps. huaardii. Evidence for the sufficient prey were available,and rapid devel- similarities in egg distribution of Ps.colurnbiae opment of this predator is of paramount im- and Ps. howardiiis confirmed by using statistics portance in an ephemeral aquatic habitat. that are sensitiveto distributional characteristics Psorophoracolumbiae and Ps. howardii could cer- JuNn, 1985 J. Ar"r.Mosq. CoNrnol Assoc. 173

tainly develop successfully in winter if egg hatch Coachella Valley, California, Ann. Entomol. Soc. was stimulated by flood irrigation since habitat Am. 57:383-387. permanence would be sufficient. However. Gunstream, S. E. and R. M. Chew. 1967. The ecology hatching could also be stimulated by normai of Psorophora confinnis (Diptera: Culicidae) in rainfall which may not be sufficient to allow southern California II. Temperature and development. Ann. Entomol. Soc. Am. 60:434-439. ,4evelopment when water temperatures are low. Harden, F. W., H. R. Hepburn and B. Ethridge, By hatching J. almost exclusively in response to 1967. A history of mosquitoes and mosquito borne crown flood irrigation, Ae. aexans is able to diseases in Mississippi, 1699-1965. Mosq. News compensate for increased development time 27:60-66. during unfavorable winter conditions, since Hill,M. N.,J.A. Chandlerand R. B. Highton.1977. habitat duration is more permanent. Com- A comparison of mosquito populations in irrigated monly seen is a l7-day development for Ae. and non-irrigated areas of Kano Plains, Nyanza Province, Kenya, pp. 307-315.1n: E. B. Wor- aexans following irrigation in January, when thington (ed.). Arid land daily water temperature in citrus furrows irrigation in developing countries: Environmental effects and problems. ranges between 4- lSC'. Pergamon Press, Inc., Elmsford, NY. Horsfall, W. R., R. J. Novak and F. L. Johnson. 1975. Aedes uexans ACKNOWLEDGMENTS as a flood-plain mosquito. Environ. Entomol: 4:675-681. Krebs, C. 1972. Ecology: The experimental I am grateful to D.B. Carlson, G. D. Dodd, J. J. analysis H. Frank, L.P. Lounibos and G.F. O'Meara for of distribution and abundance, Harper and Row, New York. their useful comments and criticism of the Mclachlan, A. and M. A. Cantrell. 1980. Survival manuscript. J. strategies in tropical rainpools. Oecologia 47:344- 351. ReferencesCited Olson, J. K. and C. L. Meek. 1977. Soil moisture conditions that are most attractive to ovipositing Al-Azawi, A. and R. M. Chew. 1959. Notes of the females of Psorophora columbiae in Texas ricelands. ecofogy of the dark field mosquito Psorophoracon- Mosq. News 37:19-26. finnis, in the CoachellaValley, California (Diptera: Portman, R.F. 1954. Conrol of mosquitoes in Culicidae).Ann. Entomol. Soc.Am. 52:345-351. California ricefields. J. Econ. Entomol. 47:818. Boyd, M. F. 1941.A historicalsketch of the preva- Pritchard, G. and P. J. Scholefield. 1983. Survival of lence of malaria in North America. Am. J. Trop. Aed,es lamae in constant area ponds in southern Med,.2l:223-224 Alberta (Diptera: Culicidae). Can. Entomol. Briscoe, M. S. 1952. The relation of insects and I l5:183- 188. insect-borne diseasesto the vegetation and envi- Reeves. W.C. and W. McD. Hammon. 1962. ronment in Liberia. Ecology 33:187-214. Epidemiology of the arthropod-borne en- Curtis,G. A. andJ. H. Frank. 1981.Establishment of cephalitides in Kern County, California, 1941- Aedesuexans in citrus groves in southeastern 1952. Univ. Calif. Publ. Public Health 4:l-256. Florida. Environ. Entomol. l0: 180-182. Shemanchuk, J.A. 1959. Mosquitoes in the irrigated Daniel, C. and F. Wood. 1980. Fitting equationsto areas of southern Alberta and their seasonal data.John Wiley and Sons,New York. changes in abundance and distribution. Can. J. Dunn, O. J. 1964.Multiple comparisonsusing rank Zool. 37:89S912. sums. Technometrics 6:241-252. Surtees, G. 1970a. Large scale irrigation and ar- Edmunds, L. R. 1958.Field observationson the habits bivorus epidemiology, Kano Plain, Kenya. J. Med. and seasonal abundance of mosquito larvae in Entomol. 7:509-517. ScottsBluff County, Nebraska (Diptera: Culicidae). Surtees, G. 1970b. Effects of irrigation on mosquito Mosq.News 18:23-26. populations and mosquito-borne diseases in man, Gunstream,S. E. and R. M. Chew. 1964.Contribution with particular reference to ricefield extension. Int. to the ecologyof Aedesaexans (Diptera: Culicidae)in J. Environ. Stud. l:35-42.