Relative Distribution,Diversity,and Bloodmeal Sources of Mosquitoes and Known Vectors of Rift Valley Fever Phlebovirus in Three Differing Ecosystems in Bura,Tana River County,Kenya

Abstract

Environmental modifications disturb the equilibrium of mosquito populations, altering the risk of mosquito-borne diseases. Mosquito distribution, diversity, and bloodmeal sources were examined to compare Rift Valley fever (RVF) risk among irrigated, riverine, and pastoral ecosystems in Bura, Tana River County, Kenya, between September 2014 and June 2015. Thirty-eight households and 21 irrigation fields were selected for the study. Mosquitoes were trapped with carbon dioxide-impregnated CDC traps, one trap per household and three traps per irrigated field, and morphologically identified using taxonomic keys. Host DNA was extracted from engorged females and cytochrome b genes amplified by PCR to identify sources of bloodmeals. A total of 21,015 mosquitoes were collected; 5742 within households in the 3 ecosystems and 15,273 within irrigated fields. Mosquitoes collected within irrigated fields belonged to 8 genera and 37 species, while those from households within the irrigation scheme belonged to 6 genera and 29 species. Collections from riverine and pastoral households belonged to five and four genera, respectively. The most abundant genera in the irrigated fields were Aedes (21%) and Mansonia (22%), while Anopheles (43%) was the most abundant within households. Most mosquitoes in riverine and pastoral households belonged to Anopheles (76%) and Aedes (65%) genera, respectively. Seasonal variation driven by rainfall was evidenced by spikes in mosquito numbers within irrigated and riverine ecosystems. Host species identification revealed that goats and humans were the main sources of bloodmeal. There was an overall increase in mosquito abundance and diversity as a result of the presence of the irrigated ecosystem in this county, and an increased availability of highly RVF-susceptible hosts as a result of the establishment and concentration of residential areas, promoting potential vector–host contacts. These results highlight the impact of anthropogenic changes on mosquito ecology, potentially heightening the risk of transmission and maintenance of RVF in this region.

Introduction

Land use and land cover changes have been directly implicated in the emergence or spread of arthropod-borne diseases (Patz et al. 2000) such as West Nile fever, malaria, and lyme disease. Mosquitoes seem particularly vulnerable to these changes, which significantly alter their population dynamics, species composition, and competence (Johnson et al. 2008). Known mosquito vectors of Rift Valley fever (RVF) Phlebovirus are broadly grouped into two. “Reservoirs” are mostly Aedes species believed to maintain the virus through transovarial transmission and subsequently lay their infected eggs in temporary “dambos” or semipermanent pools of water that form in low-lying plains after heavy rains. “Amplifying” vectors consist of mostly Culex species found within more permanent water bodies. These vectors become infected and amplify the virus, resulting in the transmission of disease to more hosts, thereby propagating the epidemic (Sang et al. 2010, Marcantonio et al. 2015).

Recent studies in Kenya have attempted to associate the abundance and geographic distribution of vectors of RVF Phlebovirus with spatio-temporal differences in RVF risk. In Sang et al. (2017), the abundance and diversity of known vectors of RVF Phlebovirus varied significantly between different ecological sites. This study aimed to examine and compare the abundance, seasonal variation, and species distribution of mosquitoes and vectors of RVF Phlebovirus in three differing ecosystems, within the context of irrigation expansion efforts in Tana River County and determine their vertebrate sources of bloodmeal. Other variables such as animal herd sizes and human hosts were also investigated. The results will help determine the likelihood of changes in RVF risk in this region because of changes in mosquito population dynamics and in formulating comprehensive strategies for disease control. This study was part of a larger one whose objective was to evaluate and compare the interepidemic transmission of RVF Phlebovirus in sheep and goats in differing ecosystems, and the risk factors precipitating these differences (Mbotha et al. 2017).

Materials and Methods

Study sites

Mosquitoes were captured in three ecologically distinct sites within Bura, Tana River County between September 2014 and June 2015 (between 1.32278S, 39.9532E and 1.12218S, 39.7046E, Fig. 1). These were Bura irrigation scheme and Husingo and Chifiri villages. Bura irrigation scheme, situated along Tana River near Bura town, constitutes large tracks of land under irrigated cultivation. Households farming these lands congregate adjacent to the fields. They also keep small herds of sheep and goats for household use. Husingo, also located near the river, represents a riverine locale. It is surrounded by forested areas, bushes, and shrubs and occasionally attracts wildlife. Inhabitants practice seasonal farming on small tracks of land that flood during the rainy season and keep small herds of sheep and goats. Chifiri is located deeper in the drier hinterland and represents a pastoral ecology. Inhabitants keep large herds of diverse livestock, including sheep, goats, cattle, camels, and donkeys for commercial purposes. There is very scarce vegetation in Chifiri.

FIG. 1.

Schematic representation of the three study sites in Bura, Tana River County, sampled between September 2014 and June 2015; Bura irrigation scheme and Husingo and Chifiri villages. The map is not drawn to scale. Original map published in Mbotha et al. (2017).

Selection of participating households

The selection of households has been described previously (Mbotha et al. 2017). In brief, 21 households in Bura irrigation scheme and 13 in Husingo village with RVF-seronegative animals were selected for sampling for the previous study, and consequently used for this study. An additional four households were randomly selected from a list of all households in Chifiri village, which was not included in the previously mentioned study. Household characteristics such as size, animal herd size, source of livelihood, and insecticide-treated bed-net use were collected.

Mosquito collections and handling

Study sites were visited in the years 2014 (September and November) and 2015 (January, March and June), respectively. Carbon dioxide-baited CDC light traps (John W. Hock, Gainesville, FL) were placed within compounds of selected households, between the house and the livestock night shed adjoining it, from 5PM to 6AM for one night per visit. They were placed about 5 feet from the ground. In the irrigation scheme, three traps were set inside selected fields. Traps set in fields were placed in different locations around flooded areas or other potential larval breeding sites. Trapped mosquitoes were transferred to a field laboratory and immobilized using 99.5% triethylamine (Sigma-Aldrich, St. Louis, MO), sorted, counted, labeled as the total number per household or field per trap, and stored in liquid nitrogen. The presence and number of larval breeding grounds were noted while immature stages of mosquitoes were collected from irrigation feeder canals, block feeders, unit drains, and stagnant or marshy water between crops. They were left to mature in the field laboratory and included with the rest. All collections were transported to the Kenya Medical Research Institute's (KEMRI) laboratory and identified using the keys of Edwards (1941) and Jupp (1986) and pooled in groups of up to 25 by collection date, species, sex, and site. Blood-fed mosquitoes were preserved individually.

Identification of source of bloodmeal

Extraction of genomic DNA from blood-fed mosquitoes was done using QIAGEN DNeasy blood and tissue kit (Qiagen GmbH, Hilden, Germany) following the manufacturer's protocol and samples stored at −20°C. Extracted DNA was used as template in PCR amplification of the targeted mitochondrial cytochrome b 358-bp region using primers L14841 (forward 5′-CCATCCAACATCTCAGCATGATGAAA-3′) and H151494 (reverse 5′-GCCCCTCAGAATGATATTTGTCCTCA-3′) (Kocher et al. 1989). Purified amplicons were sequenced in forward direction (Molecular Biology GmbH, Germany) and sequences edited in BioEdit (Hall 1999). Edited sequences were compared to nucleotide entries at the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov) via BLASTN^® (https://blast.ncbi.nlm.nih.gov/Blast.cgi) (Altschul et al. 1990). The cutoff for identity was set at 99%.

Data management and analysis

Regression analysis was done on mosquitoes collected within households in Bura and Husingo for comparison, while those from Chifiri were left out due to selection bias, as this village had not been included in the previous study (Mbotha et al. 2017). Multivariable evaluation of counts of mosquitoes collected per trap during each visit was compared using a negative binomial model (MASS package) in R 3.3.3 software (R Development Core Team 2010, Hilbe 2011, Bates et al. 2014). Two analyses were done; the first included mosquitoes from households in the two sites, while the second was limited to known vectors of RVF Phlebovirus namely Aedes mcintoshi, Aedes tricholabis, Aedes ochraceus, Anopheles squamosus, Culex poicilipes, Culex bitaeniorhynchus, Culex univittatus, Culex pipiens sl, Mansonia africana, and Mansonia uniformis. Independent variables examined included rainfall, month of visit, number of humans and animals in a household, and household livelihood. Household was incorporated as a random effect while village as a fixed effect. Bidirectional elimination of variables was used and models compared using the Akaike Information Criterion.

Ethical approval

Ethical approval was obtained from the Ethics and Scientific Review Committee (ESRC) of the Africa Medical and Research Foundation (AMREF) number (REF: AMREF-ESRC P65/2013).

Results

Relative abundance and species composition

A total of 21,015 mosquitoes and larvae were collected and examined. Of these, 5742 were from all households (Table 1), 14,498 were from irrigated fields in Bura (Table 2), and 775 were immature stages from breeding habitats within irrigated fields in Bura (Table 3). Over half (11,539) of all mosquitoes were known vectors of RVF Phlebovirus (Table 4).

Table 1.

Number of Mosquitoes Collected and Identified from Selected Households During Five Visits to Bura Irrigation Scheme and Husingo and Chifiri Villages Between September 2014 and June 2015

	Aedes spp.	Aedomyia spp.	Anopheles spp.	Culex spp.	Ficalbia spp.	Mansonia spp.	Total
Bura (irrigated) total	868	308	2187	1602	1	93	5059
September 2014	11	108	77	328	0	2	526
November 2014	202	31	25	299	0	2	559
January 2015	90	20	995	368	1	38	1512
March 2015	27	111	177	292	0	41	648
June 2015	538	38	913	315	0	10	1814
Chifiri (pastoral) total	26	0	2	11	0	1	40
September 2014	0	0	0	0	0	0	0
November 2014	26	0	0	1	0	0	27
January 2015	0	0	0	4	0	1	5
March 2015	0	0	0	3	0	0	3
June 2015	0	0	2	3	0	0	5
Husingo (Riverine) total	11	2	492	135	0	3	643
September 2014	0	0	1	14	0	0	15
November 2014	5	1	1	6	0	0	13
January 2015	0	0	393	82	0	1	476
March 2015	0	0	1	18	0	1	20
June 2015	6	1	96	15	0	1	119
Total	905 (15.8%)	310 (5.4%)	2681 (46.7%)	1748 (30.4%)	1 (0.0%)	97 (1.7%)	5742

Table 2.

Number of Mosquitoes Collected and Identified from Traps Set Within Irrigated Fields During Five Visits to Bura Irrigation Scheme Between September 2014 and June 2015

	Aedes spp.	Aedomyia spp.	Anopheles spp.	Coquillettidia spp.	Culex spp.	Ficalbia spp.	Mansonia spp.	Total
September 2014	81	461	31	1	614	0	2	1190
November 2014	871	560	48	0	734	0	26	2239
January 2015	1249	550	1409	1	1080	26	1830	6145
March 2015	183	436	218	0	720	0	736	2293
June 2015	707	368	310	0	601	0	645	2631
Total	3091 (21.3%)	2375 (16.4%)	2016 (13.9%)	2 (0,01%)	3749 (25.9%)	26 (0.18%)	3239 (22.3%)	14498

Table 3.

Number of Immature Mosquito Stages Collected from Their Breeding Habitats Within the Irrigated Fields During Five Visits to Bura Irrigation Scheme Between September 2014 and June 2015

	Arrow root holes	Block feeder	Drain feeder	Farm furrow	Feeder canal	Stagnant water	Unit canal	Unit drain	Total
Aedes (5.2%)	0	0	9	0	1	18	1	11	40
Aedomyia (1.4%)	0	0	0	0	11	0	0	0	11
Anopheles (20.3%)	1	2	42	0	26	63	9	14	157
Culex (72.8%)	21	1	182	4	128	153	14	61	564
Uranotaenia (0.39%)	0	0	0	0	3	0	0	0	3
Total	22	3	233	4	169	234	24	86	775

Table 4.

Total Number of Known Mosquito Vectors of Rift Valley Fever Virus Collected and Identified from Households and Irrigated Farms During Five Visits to Bura Irrigation Scheme and Husingo and Chifiri Villages Between September 2014 and June 2015

	Aedes (30.1%) Aedes Mcintoshi Aedes ochraceus Aedes Tricholabis	Anopheles (1.8%) Anopheles squamosus	Culex (36.8%) Culex bitaeniorhynchus Culex pipiens Culex poicilipes Culex Univittatus	Mansonia (31.3%) Mansonia africana Mansonia uniformis	Total
Bura (irrigated) total	3448	198	4156	3602	11,404
September 2014	110	3	751	5	869
November 2014	1092	0	907	28	2027
January 2015	1383	147	1272	2266	5068
March 2015	207	18	595	650	1470
June 2015	656	30	631	653	1970
Chifiri (pastoral) total	21	1	6	1	29
September 2014	0	0	0	0	0
November 2014	21	0	1	0	22
January 2015	0	0	3	1	4
March 2015	0	0	1	0	1
June 2015	0	1	1	0	2
Husingo (riverine) total	10	5	88	3	106
September 2014	0	0	3	0	3
November 2014	5	0	4	0	9
January 2015	0	2	69	1	72
March 2015	0	0	0	1	1
June 2015	5	3	12	1	21
Total	3479	204	4250	3606	11,539

Mosquitoes collected in all households in the 3 ecosystems consisted of 29 species of 6 genera namely Aedes (15.76%), Aedomyia (5.40%), Anopheles (46.69%), Culex (30.44%), Ficalbia (0.02%) and Mansonia (1.69%, Fig. 2). Members of Aedes comprised Ae. aegypti, Ae. mcintoshi, Ae. ochraceus, Aedes simpsoni, Aedes stegomyia spp., Aedes sudanensis and Ae. tricholabis Aedeomyia genus consisted of Aedeomyia furfurea and other unidentified Aedomyia spp., Anopheles genus consisted of Anopheles constani, Anopheles funestus, Anopheles gambiae, Anopheles pharoensis, An. squamosus, and unidentified Anopheles spp., Culex genus included Culex annulioris, Cx. bitaeniorhynchus, Culex ethiopicus, Cx. pipiens, Cx. poicilipes, Culex tigripes, Cx. univittatus, Culex vansomereni, and other unidentified Culex spp. One mosquito was identified as belonging to Ficalbia genus while members of Mansonia were Ma. africana and Ma. uniformis. Majority of mosquitoes (88.1%) from households were from the irrigated ecosystem. It was also the most diverse, with six genera being recorded. The riverine ecosystem had five genera, while the pastoral had four. Overall, majority of mosquitoes captured from households belonged to Anopheles and Culex genera. Univariable analysis revealed no significant difference between the number of mosquitoes from the irrigated and riverine households (IRR = 1.02, CI = 0.73–1.35, p = 0.92).

FIG. 2.

Distribution of mosquito genera collected from all households in Bura irrigated scheme, the riverine Husingo village, and the pastoral Chifiri village during five visits between September 2014 and June 2015.

Mosquitoes captured within irrigated fields in Bura exhibited the highest diversity in species composition, registering 35 species belonging to 7 genera. These were Ae. aegypti, Ae. funestus, Ae. mcintoshi, Ae. ochraceus, Ae. simpsoni, Ae. stegomyia spp., Ae. sudanensis, Ae. tricholabis, Aedes poicilipes, Aedes tarsalis, and other unidentified members. Aedomyia genus included Ad. furfurea and unidentified members. Members of Anopheles genus included An. constani, An. funestus, An. gambiae, An. pharoensis, An. squamosus, Anopheles sudanensis, and other unidentified species. Two Coquillettidia aurites mosquitoes, belonging to genus Coquillettidia were also identified. Members of Culex genus included Cx. annulioris, Culex antennatus, Cx. bitaeniorhynchus, Cx. ethiopicus, Cx. pipiens, Cx. poicilipes, Culex uniformis, Cx. univittatus, Cx. vansomereni, Culex zombaensis, and other unidentified members. Filcalbia genus members were Filcalbia splendens and other unidentified species. Members of Mansonia genus were Ma. africana, Ma. uniformis, and other unidentified species.

Majority of known vectors of RVF Phlebovirus were captured within the irrigated ecosystem (98.8%). All ecosystems, however, registered at least 1 member of each of the 10 species of known vectors. Immature stages collected from breeding habitats within irrigated fields consisted of 15 species belonging to 5 genera, namely Ae. mcintoshi, Ad. furfurea, An. funestus, An. gambiae, and An. squamosus and unidentified Anopheles spp., Cx. annulioris, Cx. bitaeniorhynchus, Cx. ethiopicus, Cx. pipiens, Cx. poicilipes, Cx. univittatus, Cx. vansomereni, and other unidentified Culex spp., and 3 members belonging to Uranotaenia genus.

In addition to increased mosquito density after the rains, there were notable differences in spatio-temporal distribution of species in all ecological sites. Anopheles species was the most abundant genus and had the highest peak in the irrigated and riverine ecosystems after the rains, while Aedes species had the highest peak in the pastoral Chifiri village. Culex and Mansonia species were collected in all sites and maintained their relative abundance throughout the study period. The most abundant genus within households in Bura and Husingo was Anopheles, whereas collections from irrigated fields in Bura mirrored the dry pastoralist ones with the abundant genera being Aedes and Mansonia.

Seasonal variation in mosquito abundance

Anopheles genus had the highest spike after the rains in the irrigated and riverine ecosystems. Culex was the only genus present in all three study sites during all visits except September 2014, and its numbers remained relatively constant within each ecosystem (Fig. 3A–C). The first visit in September 2014 accounted for the least number of mosquitoes collected overall (9.42%). Numbers increased slightly in November 2014 (10.43%) then spiked in January 2015 (34.71%), mainly caused by an increase of Anopheles spp. The March 2015 visit registered a significant dip (11.69%) that bounced back in the final June 2015 visit (33.75%). This pattern was, however, not seen in Chifiri, as there were no peaks in mosquito numbers in January and June 2015. Seasonal variation among mosquitoes from irrigated fields in Bura and known vectors of RVF Phlebovirus followed the same pattern as those from households, with majority captured after the rains in January 2015. Most larvae and pupae of mosquitoes were collected in November 2014 (54.8%) and June 2015 (26.7%).

FIG. 3.

Distribution of mosquito genera collected from households in Bura irrigated scheme (A), the riverine Husingo village (B), and the pastoral Chifiri village (C) during five visits between September 2014 and June 2015.

Results of regression analysis

There was no significant difference in counts of mosquitoes collected between the irrigated and riverine ecosystems for both total mosquitoes and known vectors of RVF Phlebovirus. There were, however, significant differences in counts of mosquitoes collected between different months for total mosquitoes (Table 5), while known vectors of RVF Phlebovirus experienced a significant reduction in numbers captured in June. Larger households tended to have significantly lower mosquito catches. Animal herd size did not have any impact on the number of mosquitoes captured (Table 6).

Table 5.

Results of a Multivariable Model for the Counts of Mosquitoes Collected from Households in Bura Irrigation Scheme and Husingo Village Only in Tana River County, Kenya, During September 2014–June 2015

	Mosquitoes collected	IRR	Confidence interval	p
Site
Irrigated	5059
Riverine	643	0.77	0.56–1.08	0.11
Month
September 14	541	0.51	0.37–0.69	<0.001^*
November 14	572	0.44	0.32–0.59	<0.001^*
January 15	1988
March 15	668	0.52	0.39–0.71	<0.001^*
June 15	1933	1.01	0.77–1.33	0.92
Hosts
Humans in households
1–5	2825
6–10	2457	1.05	0.81–1.36	0.69
>10	420	0.67	0.48–0.96	0.02^*
Animals in households
1–5	1314
6–15	3074	1.18	0.91–1.53	0.23
>15	1314	1.06	0.77–1.46	0.71

IRR, incidence risk ratio.

There was a significant difference in the number of mosquitoes collected between the dry months of September through November 2014 and March 2015, and the wet months of January and June 2015. A significant difference was also found in mosquitoes collected between household having more than 10 members compared to those with less.

Table 6.

Results of a Multivariable Model for the Counts of Known Mosquito Vectors of Rift Valley Fever Virus Collected from Households in Bura Irrigation Scheme and Husingo Village in Tana River County, Kenya, During September 2014–June 2015

	Mosquitoes collected	IRR	Confidence interval	p
Site
Irrigated	1456
Riverine	124	0.74	0.48–1.15	0.16
Month
September 14	191	0.69	0.46–1.04	0.07
November 14	351	0.88	0.61–1.28	0.05
January 15	514
March 15	283	1.03	0.71–1.53	0.88
June 15	241	0.57	0.40–0.81	0.001^*
Hosts
Humans in household
1–5	931
6–10	519	0.72	0.51–0.98	0.04^*
>10	130	0.59	0.38–0.95	0.03^*
Animals in household
1–5	393
6–15	830	1.41	1.00–1.99	0.06
>15	357	1.27	0.85–1.92	0.23

There was a significant difference in the number of known vectors of RFV virus collected in June 2015 compared to the rest of the months. A significant difference was also found in mosquitoes collected between households with more than six members compared to those with less.

Host species identification of bloodmeals

There were 343 engorged females collected overall from the 3 sites. Majority were from households (301/343), while the rest were from irrigated fields. Most engorged mosquitoes from households originated from Bura (86%), Husingo (12%), and Chifiri (2%). Most were captured a few months after the start of the short rainy season in November 2014 (18%), January 2015 (43%), and the long rainy season in June 2015 (29%). Bloodmeal analysis revealed 302 samples contained vertebrate DNA (88%), while the rest were unidentified (12%). One-third of mosquitoes (n = 100) with identified host species DNA were An. funestus, while the rest were Ae. mcintoshi (n = 46, 15%), Cx. annulioris (n = 40, 13%), Cx. univittatus (n = 30, 10%), and other species. Of these, 126 (42%) were known vectors of RVF Phlebovirus. Majority (78%) of samples contained DNA originating from goats, while 47 (16%) were of human origin (Table 7). Majority of engorged females containing unidentified host DNA were An. funestus (n = 28, 68%), Ae. mcintoshi (n = 6, 15%), An. gambiae (n = 3, 7%), two unidentified Aedes and Anopheles species, one An. squamosus, and one Cx. vansomereni.

Table 7.

Vertebrate Sources of Bloodmeals Identified from Female Fed Mosquitoes Collected From Bura Irrigation Scheme Households and Fields and Husingo and Chifiri Villages Between September 2014 and June 2015

	Antelope	Bird	Cow	Donkey	Goat	Human	Mouse	Sheep	Total
Ae. mcintoshi ^a	1	0	6	0	29	8	1	1	46
Ae. ochraceus ^a	0	0	0	0	4	0	0	0	4
Aedes Sudanensis	0	0	0	0	0	2	0	0	2
Anopheles funestus	0	0	3	0	88	8	0	1	100
Anopheles gambiae	0	0	0	0	10	6	0	0	16
An. squamosus ^a	0	0	0	0	3	0	0	0	3
Anopheles spp.	0	0	0	0	3	0	0	0	3
Culex spp.	0	1	0	0	7	2	0	0	10
Culex annulioris	0	0	0	1	32	5	0	2	40
Cx. pipiens ^a	0	0	0	0	18	3	0	2	23
Cx. poicilipes ^a	0	0	0	0	4	2	0	0	6
Cx. univittatus ^a	0	0	0	0	23	7	0	0	30
Culex vansomereni	0	1	0	0	2	0	0	0	3
Filcalbia spp.	0	0	0	0	0	2	0	0	2
Ma. africana ^a	0	0	0	0	12	2	0	0	14
Total	1	2	9	1	235	47	1	6	302

Represents known vectors of RVF Phlebovirus.

RVF, Rift Valley fever.

Discussion

This study was informed by the current global increase in the emergence and spatial expansion of infectious diseases either as a result of human actions such as land use changes, or as a consequence of human behavior such as climate change (Lindahl and Grace 2015). In particular, the spread of vector borne diseases such as RVF has been known to occur in areas favoring the emergence of massive numbers of mosquito vectors as a result of prolonged rainfall or flooding (Linthicum et al. 1985, Hightower et al. 2012). The present study investigated the impact of irrigation on the abundance, distribution and diversity of mosquitoes and potential vectors of RVF Phlebovirus during an interepidemic period between September 2014 and June 2015, in Bura, Tana River County. Multivariable analysis revealed that the irrigated and riverine ecosystems were similar in mosquito abundance and seasonality, despite one being naturally occurring, while the other man-made. Furthermore, the irrigated ecosystem was the most diverse and maintained an abundance of mosquitoes throughout all seasons. The presence of the irrigation scheme at this site led to an overall increase in mosquito population and species diversity in this region. This effect has been demonstrated in previous studies in Kenya and elsewhere, in which mosquito abundance and diversity was highly site-dependent (Lutomiah et al. 2013, Serpa et al. 2013). Imbahale et al. (2011) also indicated that most larvae breeding grounds such as feeder and drainage canals, tire tracks, and rice paddies were man-made, compared to naturally occurring habitats such as riverine flooded edges and swampy fringes.

Seasonal variation correlating with rainfall, also significantly contributed to increased mosquito numbers, especially in Bura and Husingo. Mosquito numbers peaked in January 2015, ∼6 weeks after the start of the short wet season in November 2014, due to the growth of abundant vegetation optimal for the survival of large numbers of emerging mosquitoes (Anyamba et al. 2009). This was also observed in previous studies, in which mosquito numbers correlated with weekly or monthly rainfall intensity (Patz et al. 1998, Bomblies 2012). Chifiri had virtually no breeding grounds and had significantly reduced vegetation that could support survival of large numbers of emerging mosquitoes. This was probably due to insufficient rainfall.

Although vector density is important in pathogen transmission, species distribution is equally important due to the differing epidemiological roles played by each species, all contributing to the emergence or maintenance of disease. The spatial distribution of mosquitoes in these ecosystems conforms with previous findings that showed Anopheles species, being well established in the Tana Delta (Lutomiah et al. 2013), was the most abundant in the irrigated and riverine ecosystems. Species of the flood water Aedes genus were predominantly found in the dry and arid Garissa region of north-east Kenya, a pastoral area belonging to the same ecological group as Chifiri (Arum et al. 2015). Lutomiah et al. (2013) showed that members of Mansonia were mainly collected in large swampy marshes in Baringo, Kisumu, and Budalangi in western Kenya, where they appeared to be well adapted to the hot and humid ecology with swampy breeding sites. Members of Culex have previously been found to be ubiquitous in Kenya, with different regions being inhabited by various predominant species (Sang et al. 2010).

Interestingly, more than half of all mosquitoes captured were known vectors of RVF Phlebovirus. While rainfall appeared to have an impact on seasonal variation of the overall number of mosquitoes collected within households, regression analysis revealed a significant reduction in the number of known vectors of RVF Phlebovirus captured in June, perhaps due to the shutting down of secondary and tertiary irrigation channels in most fields at the end of the growing season in April, to allow for maturation and harvesting of crops later in the year. Thus, the rain that fell later may not have been sufficient for flooding and emergence of more floodwater mosquitoes.

Large households with many children were few and tended to have more animals within the homestead. They also contained several treated mosquito nets, which could explain the reduced number of mosquitoes captured. While it may be impossible to eliminate all stagnant water especially within the irrigated ecosystem, increased distribution of insecticide-treated bed nets and use of window and door screens to limit human contact as well as other practical chemical and biological control efforts could be explored.

While most engorged mosquitoes were captured within households in the irrigation scheme, majority of the unfed mosquitoes were collected within irrigated fields. Mosquito flight range, duration, and speed have long been known to contribute to overall vector capacity in disease transmission. Host species identification revealed a limited diversity of hosts available in Bura and Husingo households, who mainly kept few goats and sheep for household use. The proportion of human bloodmeals was higher than one previous study that found it ranging between 5.1% and 5.3% in this region (Lutomiah et al. 2014). These results have significant implications for RVF Phlebovirus transmission, bearing in mind that goats and sheep are the most susceptible hosts for the virus (Chevalier et al. 2010, Lichoti et al. 2014). Furthermore, they also reveal the potential for human exposure to the virus, as emerging vectors look for alternate sources of bloodmeal, such as humans (LaBeaud et al. 2008, Mbotha et al. 2017).

Conclusion

Mosquito collections from Bura irrigation scheme and Husingo and Chifiri villages representing irrigated, riverine, and pastoral ecosystems during five visits between September 2014 and June 2015 revealed that the irrigation ecosystem is promoting mosquito abundance and diversity in this region, because of the availability of breeding grounds and resting places as well as availability of vertebrate hosts for bloodmeals. These sites might enhance RVF Phlebovirus endemicity through sustained breeding and prolonged lifespans, increasing mosquito abundance, including known vectors of RVF. The establishment or expansion of more irrigation schemes may also result in the permanent change of predominant host species available to include highly susceptible sheep and goats, while the settlement and concentration of residential areas will increase mosquito-livestock-human contact, potentially leading to human exposure. These findings highlight the impact of agricultural and other anthropogenic land use changes on mosquito ecology within the context of other natural environmental forces.

Footnotes

Acknowledgments

The authors thank Dunston Beti, John Gachoya, Reuben Lugalia, Gilbert Rotich, Phillip Tunge, and Betty Chelangat from the Kenya Medical Research Institute (KEMRI), for their technical assistance in mosquito collection and identification.

Author Disclosure Statement

No conflicting financial interests exist.

Funding Information

The study was financially supported by the International Livestock Research Institute (ILRI) through the Dynamic Drivers of Disease in Africa Consortium (DDDAC) and the CGIAR research program; Agriculture for Nutrition and Health in Nairobi, Kenya, and the German Academic Exchange Service (DAAD) in Bonn, Germany.

References

Altschul

, Gish

, Miller

, Myers

, et al. Basic local alignment search tool. J Mol Biol, 1990; 215:403–410.

Anyamba

, Chretien

J-P

, Small

, Tucker

, et al. Prediction of a Rift Valley fever outbreak. Proc Natl Acad Sci USA, 2009; 106:955–959.

Arum

, Weldon

, Orindi

, Landmann

, et al. Distribution and diversity of the vectors of Rift Valley fever along the livestock movement routes in the northeastern and coastal regions of Kenya. Parasit Vectors, 2015; 8:294.

Bates

, Mächler

, Bolker

, Walker

. Fitting linear mixed-effects models using lme4. arXiv, 2014; 67:arXiv:1406.5823.

Bomblies

. Modeling the role of rainfall patterns in seasonal malaria transmission. Clim Change, 2012; 112:673–685.

Chevalier

, Pépin

, Plée

, Lancelot

. Rift Valley fever—A threat for Europe?. Euro Surveil, 2010; 15:19506.

Edwards

. Mosquitoes of the Ethiopian Region. III. Culicine Adults and Pupae. London: British Museum (Natural History), 1941.

Hall

. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser, 1999; 41:95–98.

Hightower

, Kinkade

, Nguku

, Anyangu

, et al. Relationship of climate, geography, and geology to the incidence of Rift Valley fever in Kenya during the 2006–2007 outbreak. Am J Trop Med Hyg, 2012; 86:373–380.

10.

Hilbe

. Negative Binomial Regression. New York: Cambridge University Press, 2011.

11.

Imbahale

, Paaijmans

, Mukabana

, van Lammeren

, et al. A longitudinal study on Anopheles mosquito larval abundance in distinct geographical and environmental settings in western Kenya. Malaria J, 2011; 10:81.

12.

Johnson

, Gómez

, Pinedo-Vasquez

. Land use and mosquito diversity in the Peruvian Amazon. J Med Entomol, 2008; 45:1023–1030.

13.

Jupp

. Mosquitoes of Southern Africa, vol. 156. South Africa: Ecogilde, 1986.

14.

Kocher

, Thomas

, Meyer

, Edwards

, et al. Dynamics of mitochondrial DNA evolution in animals: Amplification and sequencing with conserved primers. Proc Natl Acad Sci USA, 1989; 86:6196–6200.

15.

LaBeaud

, Muchiri

, Ndzovu

, Mwanje

, et al. Interepidemic Rift Valley fever virus seropositivity, northeastern Kenya. Emerg Infect Dis, 2008; 14:1240–1246.

16.

Lichoti

, Kihara

, Oriko

, Okutoyi

, et al. Detection of rift valley fever virus interepidemic activity in some hotspot areas of Kenya by sentinel animal surveillance, 2009–2012. Vet Med Int, 2014; 2014:9.

17.

Lindahl

, Grace

. The consequences of human actions on risks for infectious diseases: A review. Infect Ecol Epidemiol, 2015; 5:30048.

18.

Linthicum

, Davies

, Kairo

, Bailey

. Rift Valley fever virus (family Bunyaviridae, genus Phlebovirus). Isolations from Diptera collected during an inter-epizootic period in Kenya. J Hyg, 1985; 95:197–209.

19.

Lutomiah

, Bast

, Clark

, Richardson

, et al. Abundance, diversity, and distribution of mosquito vectors in selected ecological regions of Kenya: Public health implications. J Vector Ecol, 2013; 38:134–142.

20.

Lutomiah

, Omondi

, Masiga

, Mutai

, et al. Blood meal analysis and virus detection in blood-fed mosquitoes collected during the 2006–2007 rift valley fever outbreak in Kenya. Vector Borne Zoonotic Dis, 2014; 14:656–664.

21.

Marcantonio

, Rizzoli

, Metz

, Rosa

, et al. Identifying the environmental conditions favouring West Nile Virus outbreaks in Europe. PLoS One, 2015; 10:e0121158.

22.

Mbotha

, Bett

, Kairu-Wanyoike

, Grace

, et al. Inter-epidemic Rift Valley fever virus seroconversions in an irrigation scheme in Bura, south-east Kenya. Transbound Emerg Dis, 2018; 65:e55–e62.

23.

Patz

, Graczyk

, Geller

, Vittor

. Effects of environmental change on emerging parasitic diseases. Int J Parasitol, 2000; 30:1395–1406.

24.

Patz

, Strzepek

, Lele

, Hedden

, et al. Predicting key malaria transmission factors, biting and entomological inoculation rates, using modelled soil moisture in Kenya. Trop Med Int Health, 1998; 3:818–827.

25.

R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing, 2010.

26.

Sang

, Arum

, Chepkorir

, Mosomtai

, et al. Distribution and abundance of key vectors of Rift Valley fever and other arboviruses in two ecologically distinct counties in Kenya. PLoS Negl Trop Dis, 2017; 11:e0005341.

27.

Sang

, Kioko

, Lutomiah

, Warigia

, et al. Rift Valley fever virus epidemic in Kenya, 2006/2007: The entomologic investigations. Am J Trop Med Hyg, 2010; 83(2 Suppl):28–37.

28.

Serpa

, Monteiro Marques

, de Lima

, Voltolini

, et al. Study of the distribution and abundance of the eggs of Aedes aegypti and Aedes albopictus according to the habitat and meteorological variables, municipality of São Sebastião, São Paulo State, Brazil. Parasit Vectors, 2013; 6:321.