Characterization of larval habitats of Anopheles albimanus, Anopheles pseudopunctipennis, Anopheles punctimacula, and Anopheles oswaldoi s.l. populations in lowland and highland Ecuador

Recent collection data indicate that at least four potential malaria vectors occupy more widespread distributions within the Andean highlands than in the past. Since habitat elimination is an important aspect of malaria control, it is vital to characterize larval habitats for Anopheles species within both lowland and highland sites. To that end, 276 sites within Ecuador were surveyed between 2008 and 2010. Characteristics of Anopheles-present sites for four species were compared to Anopheles-absent sites within the same geographical range and also to Anopheles-absent sites within a highland range representing potential future habitats. Thermochron iButtons© were used to describe the daily temperature variation within a subset of potential habitats. Anopheles albimanus (W.) was positively associated with permanent habitats, sand substrates, floating algae (cyanobacterial mats), and warmer temperatures in both comparisons. Anopheles pseudopunctipennis (T.) was associated with floating algae (cyanobacterial mats), warmer temperatures, and higher water clarity in both comparisons. Anopheles punctimacula (D.&K.) was negatively associated with floating algae and positively associated with dissolved oxygen in both comparisons. Anopheles oswaldoi s.l. (P.) was not significantly associated with any parameters more often than expected given larval-absent sites. The results indicate that minimum water temperatures might limit the upper altitudinal distribution of An. albimanus (18.7° C) and An. pseudopunctipennis (16.0° C). Medical professionals and scientists are increasingly concerned about the presence of malaria vectors and malaria parasites in previously malaria-free highland regions (Marten et al. 1999, Epstein 2001, Reiter 2001, Moreno 2006). Although many review articles have linked future highland malaria to climate change, only a few studies in Africa provide evidence for a meteorological cause for highland malaria (Loevinsohn 1994, Lindblade et al. 1999), while other reasons provided for the phenomenon include drug resistance, changes to vector control methodology, and human land use alterations in highland regions (Bødker et al. 2000, Shanks et al. 2000, Hay et al. 2002, Afrane et al. 2006). In the South American Andes, Anopheles pseudopunctipennis Theobald is a known malaria vector residing in highland valleys, collected frequently prior to 1950 (Levi-Castillo 1945). Recently, An. pseudopunctipennis has been implicated in Plasmodium vivax malaria transmission at an altitude of 2,300 m in Bolivia (Rutar et al. 2004). In another recent study, the authors highlighted new, widespread highland distributions in Ecuador of three other common anopheline species: Anopheles albimanus Wiedemann, Anopheles punctimacula Dyar & Knab, and Anopheles oswaldoi s.l. Perryassú (Pinault and Hunter 2011). It is therefore important to characterize the larval habitat of these four species within their current range, as well as within potential future highland habitats. An. albimanus is considered a coastal vector occurring below 500 m in altitude throughout Central and South America (Rubio-Palis and Zimmerman 1997). The habitats of An. albimanus have been well described for Central American populations. In the wet season, larvae are known to occupy rivers, marshes, estuaries, irrigation canals, and rainwater pools, whereas in the dry season, larvae are restricted to permanent habitats such as marshes (Breeland 1972, Rejmankova et al. 1991, Rejmankova et al. 1995, Roberts et al. 2002). In Mexico, larvae have been collected from transitional pastures (Rodriguez et al. 1993). Larvae are positively associated with floating cyanobacterial (blue-green algae) mats, which provide a warmer minimum temperature and lower dissolved oxygen level in the evening (Rejmankova et al. 1995, Rejmankova et al. 1996, Grieco et al. 2006) and also attract female oviposition through volatile compounds (Rejmankova et al. 2005). An. albimanus larvae have been variably associated with macrophyte vegetation in the literature (Meyer 1947, Rodriguez et al. 1993, Grieco et al. 2006), including positive associations with mangrove forests, naiads (Najas spp.), water hyacinths (Eichnornia spp.) (Rodriguez et al. 1993), submersed macrophytes (Rejmankova et al. 1993), and the floating fern (Salvinia matans) (Meyer 1947). Much less is known regarding South American populations of An. albimanus. In Peru, larvae have been collected from rice fields and drains (Guthmann et al. 2002), and in Colombia, larvae are associated with sun-exposed sites with a temperature range of 27.5 – 30.0° C (Marten et al. 1996). An. pseudopunctipennis is a coastal and piedmont malaria vector distributed widely across Central and South America (Rubio-Palis and Zimmerman 1997). In two studies of An. pseudopunctipennis habitat across its entire range, Manguin et al. (1996) and Rueda et al. (2004) observed that larvae were associated with drier environments (or frequent droughts), agricultural lands, stream pools, rocky substrates, and clear, shallow, sun-exposed waters with filamentous green algae. In Belize, larvae were associated with greater altitudes and shallower water depth than other species, as well as the presence of filamentous algae (Rejmankova et al. 1993). An. punctimacula is a secondary coastal vector of Central and South America (Rubio-Palis and Zimmerman 1997). In Belize, An. punctimacula larvae have been positively associated with macrophyte vegetation cover, detritus debris, shallow depth, and partial shade (Rejmankova et al. 1998). An. oswaldoi is a species complex distributed throughout Amazonia (Lounibos and Conn 2000, Monteiro de Barros et al. 2007). Larvae occupy both temporary and permanent pools in Venezuela (Rojas et al. 1992) and clean forest pools in Brazil (Monteiro de Barros et al. 2007). Female mosquitoes are known to selectively oviposit in habitats to avoid predation by specific predator types (Stav et al. 1999, Spencer et al. 2002) or to place larvae in close proximity to preferred food items (Reiskind et al. 2009) or in habitats suitable for temperature requirements (Rejmankova et al. 1996, Rejmankova et al. 2005). We assume that females tend to choose oviposition sites that are favorable for larval development. However, it is reasonable to predict that at the edge of a species’ range, such as at very high elevations, females might be provided with fewer site choices and therefore suitability might be far more variable (Koenraadt et al. 2006). Therefore, we made an effort to incorporate all possible larval habitats that we encountered within the entire current range of these species in Ecuador, including both highland and lowland sites. Since most studies of larval habitat characterization of these species have been conducted in Central America (Rejmankova et al. 1993), we expect available larval habitats for South American populations to be sufficiently different to warrant investigation. Many of the South American studies (Manguin et al. 1996, Marten et al. 1996) only describe sites where Anopheles are present, which could be biased by the actual proportion of these parameters in all existing potential habitats. The presence of a species within a specific type of habitat might be a factor of availability rather than suitability. Therefore, we compare Anopheles-present sites to Anopheles-absent sites within the current range of each species. We also compare Anopheles-present sites to highland Anopheles-absent sites that could serve as future larval habitats. For each species, the daily minimum, average, and maximum temperatures were determined and future maximum altitudes based on daily minimum temperatures were estimated, since other authors cite the importance of daily temperature fluctuations for anopheline larval habitat suitability (Huang et al. 2006, Paaijmans et al. 2008). Habitat data and Anopheles larvae (when present) were collected from 276 sites (Figure 1) encompassing a wide variety of habitat types in Ecuador, from September–December, 2008 and August–December in 2009 and 2010. We sampled all potential habitats encountered, characterized by a stationary water surface, including river edges clogged with vegetation. At each site, we sampled for Anopheles larvae by skimming the water surface a standardized 30 times, using a 13 cm diameter plastic mosquito dipper (BioQuip). For each dip, the total number of Anopheles larvae and non-Anopheles mosquito larvae were recorded. Anopheles larvae were individually collected from the dipper with a pipette and placed in 95% ethanol for transportation to the laboratory for subsequent identification. Map of all study sites sampled in 2008, 2009, and 2010 (black and red circles) in Ecuador. As an example of larval absent site selection, An. punctimacula-positive sites are indicated with red circles, absent sites (•) within the species’ current distribution are outlined with a thick dotted line, and absent sites (•) within the highland distribution are outlined with a dashed line. Inset: Map of Thermochron iButton® deployment sites (▴). In both maps, the approximate extent of the Andes is shaded in green. Maps made in ArcGIS v. 9.3 (ESRI). Late-instar Anopheles larvae were identified morphologically using the key of Gorham et al. (1973) with Faran (1979). Larvae too small to be identified morphologically were compared to a subsample of large, identifiable larvae using molecular comparison of the 800 base pair region of the mitochondrial cytochrome oxidase 1 gene (CO1) as in Pinault and Hunter (2011). Specimens were submitted to the Ecuadorian National Collection (PUCE, Quito) and genetic sequences were submitted to GenBank (accession numbers JN412826-JN412843). Data recorded at each site included the following large-scale factors: latitude, longitude, and altitude, water body type (nine categories), primary human site use (seven categories), presence or absence of pollution or contamination (e.g., detergent, oil, garbage), general ecosystem type (eight categories), and water bodies were categorized as natural vs anthropogenic and permanent vs temporary. Abiotic factors that were recorded included: pond surface area and depth where the larvae were residing (using a measuring tape), water temperature, pH and conductivity (Hanna Instruments HI98129 combination water tester), water clarity and incident sunlight (10% classes from visual estimates), and dominant substrate (eight categories). Although we also measured air temperature, we only used surface water temperature in the analysis since air temperature has been shown to be a poor indicator of habitat suitability for Anopheles species (Paaijmans et al. 2008). Dissolved oxygen was estimated using the following equations: Where DO= dissolved oxygen in mg/L, P= barometric pressure (torr), calculated from altitude, p= vapor pressure (torr), s= salinity correction factor, t= temperature (measured) (Weiss 1970). Biotic factors that were recorded included woody debris and detritus debris (present/absent), vegetation or absence of and floating and of or and of observed in 30 Since all temperature in study sites were the the daily temperature we a total of sites to sites for each Anopheles species and other for daily temperature daily average, and maximum At each site, a Thermochron iButton® Data was in an with all air and to on the surface of the water on the edge of the where larvae were were to a placed at the of the In a study, the were to have a very small of C to C) were measured at for at least two and minimum, average, and maximum temperatures determined for each Larvae of An. albimanus with An. pseudopunctipennis in sites and were other we the data at the level rather than at the Since many studies only characterize habitats used by Anopheles larvae (Marten et al. 1996, Manguin et al. and might therefore be biased by the at which these parameters in all potential habitats, the of parameters observed in Anopheles-present sites were compared to the of parameters in larval-absent to in habitat to the actual habitat of were to compare sites to larval-absent sites for each species. to current larval habitats, we compared sites to larval-absent sites within the current distribution of the species. Although many sites were sampled throughout Ecuador, we sites that were for habitat availability within the current species In the of absent sites for the we sites that were on the same of the Andes as the species’ and absent sites were from the analysis were m higher than the species’ altitudinal sites for An. albimanus and An. pseudopunctipennis in the of were from analysis since these species are not distributed in these general regions (Pinault and Hunter 2011). to the between current habitat and potential future habitat in highland sites were compared to highland larval-absent sites. were as highland were m on the and 800 m on the of the Andes, since these sites were all in of As absent sites were sampled on the same of the Andes as the species’ and sites at altitudes higher than m were not since the altitude for Anopheles larvae recorded in Pinault and Hunter was example of site for larval-absent sites for An. punctimacula is provided in For each species and of were used to compare natural vs permanent vs and of oil, detergent, vs not and of dominant substrates, woody debris and detritus debris, between observed sites of Anopheles-present and expected sites from the of Anopheles-absent Since were we used a correction for which the level of by where is the number of in the and the of positive results and Therefore, for the results were considered to water body type and human land use associated with each species are described rather than since were too many observed for each to For both current distribution sites and potential future highland algae type and macrophyte vegetation type or absence of four were using a where the variable was the presence or absence of a given mosquito species. To which factors temperature, dissolved water incident surface depth, and to the presence or absence of a species in a larval a was used on data using the for data and was included as an factor for sites only within the species’ current average, and maximum daily temperatures from Thermochron were to The of was determined for each of the temperature An. albimanus were collected from the wet forest which with from the higher proportion of larvae was collected than expected in permanent habitats An. albimanus larvae were often associated with human and (Figure Larvae were collected from a variety of water body including rice marshes, and river edges (Figure An. albimanus larvae were associated more often with sand than expected from potential habitat availability was of larvae with woody debris, detritus debris, or Larvae were positively associated with the presence of floating which included cyanobacterial mats, with of sites positive for floating algae Larvae were not significantly associated with the presence of different types of macrophyte vegetation. temperature and conductivity were factors in the An. albimanus larvae were associated with warmer temperatures and higher conductivity than the to larval of higher salinity in coastal sites of human land on sites where An. albimanus An. pseudopunctipennis An. punctimacula and An. oswaldoi larvae were collected in 2008, 2009, and for dominant human land use at a given were categorized from to land use human and body types where An. albimanus An. pseudopunctipennis An. punctimacula and An. oswaldoi larvae were collected between in Ecuador, as a body types were categorized river edge pond rice irrigation and An. pseudopunctipennis larvae were collected from in in the wet valleys, and a Larvae were collected more often human and in an human land use (Figure An. pseudopunctipennis larvae were collected more often in and river edges than in other water body types (Figure all the of the factors were As with An. An. pseudopunctipennis larvae were positively associated with floating algae of as well as negatively associated with the presence of macrophyte vegetation of water temperature, and water clarity were factors in the larvae were associated with higher altitudes warmer temperatures C) and higher water clarity than expected from Anopheles-absent sites within their distribution An. punctimacula larvae were collected in the wet and regions of Ecuador. Larvae were associated with and (Figure and were collected from and river edges (Figure of the were An. punctimacula larvae were not significantly associated with macrophyte vegetation were negatively associated with the presence of floating algae Dissolved oxygen and incident sunlight were factors in the with larvae associated with greater estimated dissolved oxygen and less incident sunlight than expected from larval-absent as were collected from shaded pools An. oswaldoi larvae were collected in the a few sites were in of An. oswaldoi sites were associated with (Figure and were most often in and (Figure Larvae were collected more often than expected of in permanent habitats An. oswaldoi larvae were associated with the presence of floating algae were associations between larvae and algae or macrophyte vegetation the not provide any factors to the presence vs absence of An. oswaldoi within its current An. sites were compared to expected from highland absent larvae were positively associated with natural and permanent water bodies more often than expected from larval-absent sites As larvae were positively associated with sand these were less common in potential highland habitats An. albimanus larvae were more often associated with floating algae of negatively associated with algae of and positively associated with vegetation of compared to sites in highland The identified water temperature and dissolved oxygen as factors with larvae associated with warmer temperatures C) and higher of dissolved oxygen than be in potential highland habitats An. pseudopunctipennis larvae were associated with natural habitats and more often than expected from potential highland habitats Larvae were positively associated with floating algae of more often than estimated from highland sites were not significantly associated with macrophyte vegetation. The identified by the were water temperature, dissolved and water clarity with larvae associated with warmer temperatures higher conductivity higher calculated dissolved oxygen and water than available in highland sites. An. punctimacula larval habitats not significantly from potential highland habitats with to any of the measured algae or vegetation In the only water temperature and dissolved oxygen were identified as factors with larvae associated with warmer temperatures C) and higher dissolved oxygen than expected comparison to potential highland sites. An. oswaldoi larvae were not significantly associated with any of the comparison or macrophyte vegetation. As with An. only temperature and dissolved oxygen were identified as factors in the with larvae associated with warmer temperatures C) and lower dissolved oxygen than expected from highland larval-absent sites. average, and minimum temperatures from Thermochron are in with An. albimanus in the habitats, on average, and An. punctimacula in the The of the have only temperatures in by is not expected or shaded the temperatures were variable sites and not significantly with higher altitudes (Figure The between and minimum temperatures and altitude was between maximum temperature temperature and minimum temperature and altitude recorded at by Thermochron at the of current and potential Anopheles larval habitats, a in Ecuador in and 2010 of and are provided for each temperature Since minimum temperature is to larval to lower it is therefore important to it for each species. Therefore, were a small number of recorded minimum temperatures for each species, we also estimated minimum temperature using the recent maximum altitude recorded for each species (Pinault and Hunter The of the between minimum water temperature and altitude was using a of to include all data as an of the upper limit for minimum temperatures and high The minimum temperature was used to approximate maximum altitude for a 1 and C habitat on minimum temperatures a small have a on species than on at higher altitudes The study has identified habitat that are associated more often with anopheline larvae than given the range of possible habitats available within both the species’ current distributions and within a highland An. albimanus and An. pseudopunctipennis larvae were to warmer habitat temperatures in both Although an with warmer temperatures be expected comparison to highland sites the absent sites are at higher larvae were more often in warmer temperatures than available within the current range of the species. that An. albimanus and An. pseudopunctipennis be restricted by temperature to specific altitudes at Although An. pseudopunctipennis is associated with warmer temperatures, the species to be more of minimum temperatures than the other species C). the recent maximum recorded altitude is the approximate altitude limit based on minimum temperature is m for C and m for An. pseudopunctipennis in Ecuador, many within approximate altitudinal range (Levi-Castillo 1945). An. albimanus larvae were associated with floating algae in both a that the studies conducted on populations in Central America (Rejmankova et al. 1995, Rejmankova et al. 1996, Grieco et al. 2006). floating algae a in (Rejmankova et al. 1995, Rejmankova et al. 1996, Grieco et al. 2006), and is an important habitat for the larvae of species throughout its very However, the in that was associated with filamentous algae (Rejmankova et al. 1993), the An. pseudopunctipennis in Ecuador was significantly associated with floating algae. could An. pseudopunctipennis larvae also use floating algae for the of as an to their more widespread distribution than in In both An. albimanus was positively associated with permanent habitats, as well as positively associated with substrates, which were both less common in highland habitats than in lowland habitats. However, the presence of An. albimanus larvae in water bodies of different dominant that it is to regions where are In larvae were collected from human and most often in rice marshes, and river highland rice is not in Ecuador, is an land use in highland regions et al. and to be in future studies as a potential of highland An. albimanus larval Central American populations of the species have also been associated with (Rodriguez et al. 1993). An. punctimacula larvae were not significantly associated with many habitat within their current were negatively associated with floating algae. is in partial to the study by Rejmankova et al. where populations of larvae in were associated with detritus debris and shallower depth through the use of to larval habitats in Belize, An. punctimacula larvae in Ecuador were associated with shaded habitats (Rejmankova et al. 1998). Larvae were collected on in and river all of which are habitats also available in highland It is therefore that punctimacula presence is in by the availability of shaded habitat many of the human land which often The identified calculated dissolved oxygen as a factor in both for An. dissolved oxygen is and negatively with altitude, therefore An. punctimacula be from at high altitudes by of dissolved for An. punctimacula habitat where calculated was very all other sites calculated greater than studies are to by of dissolved oxygen as a possible habitat of An. punctimacula and to the between altitude and actual dissolved oxygen An. oswaldoi habitats were more to to the wide variety of in potential habitats available on the of the In An. oswaldoi larvae were often associated with and collected from and all habitats widely available in highland in a Anopheles was collected from a site at m in from known An. oswaldoi the was not identifiable to species and therefore not included in of the of study was the in a of sites that be suitable for a analysis of the many different water body types and human land available categories). As the small number of sites available in highland regions made it to compare highland Anopheles-present sites to highland Anopheles-absent sites. However, we that the current effort of 276 sites in Ecuador a within the of a small that a number of sites for was that we not in which has been shown to significantly habitat by An. albimanus and An. larvae in (Rejmankova et al. the of Ecuador, the wet from and the dry from whereas in other regions of Ecuador, the of the dry is We were to the dry to the in the of the many not to the are by frequent and of study was that we not the availability of for mosquitoes a However, effort was to regions of Ecuador, and place in regions of the As a we not the presence of as a factor, and are often used as natural and control of larvae (Rojas et al. 2004). We the presence of in all habitats present, and we not any absence of larvae associated with In most habitats with present, larvae were associated with debris, or the edge of the and were for to and As of highland regions by malaria vectors it be vital to characterize habitat by Anopheles species. Since larval habitat in highland regions is more than on the elimination of highland habitats be a malaria control mosquito and The study has determined habitat parameters that and to suitable Anopheles larval habitat the of potential habitats that in Ecuador. In or water bodies with floating algae indicate suitable habitats for An. albimanus and An. pools floating on the other be more suitable for An. punctimacula of the presence of larvae could be using larval prior to habitat elimination or in to anopheline populations in the The authors and for in Ecuador, and for laboratory for laboratory in Ecuador, for with and and two at the for the of the the following from the of as well as as study was by an to an to and an to