The study was conducted at Arjo-Didessa sugarcane irrigation scheme and its surroundings located at 395 km southwest of the capital, Addis Ababa, Ethiopia (Fig. 1). Six study clusters were randomly selected out of 15 clusters from three districts: Jimma Arjo district (Abote Didessa), Bedele District (Command 5, Bildema Deru and Ambelta) and Dabo Hana district (Kerka and Sefera Tabiya). They were selected on the basis of their proximity to the irrigation activities. The irrigation clusters (within and about 3 km from the irrigation area) were Command-5, Kerka and Abote-Didessa while the non-irrigated (4-10 km from the irrigation area) were Ambelta, Bildema Deru and Sefera Tabiya. The shortest distance between the irrigated and the non-irrigated clusters was about 4-5 km and selection of the study clusters was by assuming an average Anopheles mosquito flight range of 3 km to control overlap/contamination of mosquitoes flying from the irrigated to the non-irrigated and vice versa. A cluster is defined as an area that has 150-200 households. Both clusters had similar eco-topography. Entomological surveys were conducted from January 2018 to August 2019 in four seasons: two wet seasons and two dry seasons in the six clusters.
The districts have a total population of 215,288 and the study clusters population was 50,000. The great majority of the population depends on subsistence farming. People in the non-irrigated clusters (Ambelta, Bildema Deru, Sefera Tabiya) commonly raise cattle and cultivate mixed crops and cereals, including sorghum, rice, corn/maize, peanut and vegetables during the rainy season. Among the irrigated clusters, Command 5 is at the centre of sugarcane irrigation where farm employees live, while residents of Kerka and Abote-Didessa border the sugarcane irrigation area (Fig. 1), and use mixed farming, often planting sugarcane in their backyards. The altitude of the area ranges from 1,300 to 2,280 m above sea level with mean annual rainfall of 1,477 mm. The irrigation area and its surroundings are known to be malarious . It was formerly a wildlife sanctuary (Didessa wildlife sanctuary), but since 2006 changed to a state-owned sugarcane plantation development to supply the sugar factory. It is one of the biggest sugar development projects in Ethiopia, covering about 5,000 ha of land with future expansion plan of 80,000 ha.
Mosquito sampling and processing
Adult Anopheles mosquitoes were collected using standard Centers for Disease Prevention and Control (CDC) light traps (Model: John W. Hock CDC Light trap 512, USA) from eight randomly selected houses in each of the six clusters. At each sampling night, 16 CDC light traps were installed in each cluster. Eight light traps were placed indoors inside bedrooms at about 1.5 m above the floor near the foot end of a person sleeping under long-lasting insecticide-treated net and another eight installed outdoors at about 5 m from the same house used for indoor collection. The traps were kept running from 18:00 to 06:00 hours. A total of 192 trapping nights were spent indoors and outdoors in each cluster during the study period.
After 06:00, the CDC light traps were labelled with identifier, collected and transported to the field laboratory for processing. Live and dead mosquitoes were retrieved by mechanical aspirator from collection bags and live mosquitoes were killed using chloroform (99.8% Trichloromethane). Female Anopheles mosquitoes were sorted and identified morphologically under dissecting microscope to species using standard key . Abdominal status of the mosquitoes was determined under dissecting microscope as unfed, freshly fed, half-gravid, or gravid. Culicine and male anopheline mosquitoes were also retrieved by aspirator from the bags, counted and recorded. Each female Anopheles mosquito was preserved individually in labelled Eppendorf tube over silica gel and stored for further processing. Sample processing was done at Arjo-Didessa International Centre of Excellence for Malaria Research (ICEMR) Laboratory, Ethiopia.
Identification of Anopheles gambiae complex species
Among the total 1,418 An. gambiae sensu lato (s.l.) collected during the survey, some 225 (~16%) were randomly selected and identified to species by using species-specific polymerase chain reaction (PCR) assay at the Molecular Biology Laboratory of Tropical and Infectious Diseases Research Centre (TIDRC), Jimma University, Ethiopia. Briefly, genomic DNA was extracted using DNA extraction kit (Qiagen, Sigma Aldrich, USA) from legs and wings of each mosquito. PCR assay was carried out according to the methods of Scott et al.  using species specific primers. After PCR amplification was complete, the amplicon was loaded on 1.5% agarose gel stained with ethidium bromide and run for gel electrophoresis. Anopheles arabiensis from Sekoru insectary colony of Jimma University was used as a positive control.
Data entry and analysis was made using Microsoft Excel (Version 2016, Microsoft Corp, USA) and IBM SPSS version 20.0 (SPSS Inc., Chicago, IL, USA) statistical software packages and had been summarized with frequencies (n) and percentages (%) by species, season and irrigation levels. Chi-square (χ2) test was used to compare mosquito variation by irrigation level and season and the test was assumed significant at a p-value of less than 0.05. Indoor and outdoor mosquito density for each species per household was calculated as:
“D = n/trap-night”
where ‘D’ is density for individual mosquito species and ‘n’ is the number of mosquitoes for every species, ‘trap-night’ represents the trapping night spent in each house of all clusters. Note that the frequency of collection, the number of traps used and the number of nights spent in each season and in each cluster was similar.
Shannon diversity index was calculated to compare species richness and diversity in the irrigated and non-irrigated clusters. Shannon’s diversity index (H) was determined as follows:
Where – pi is proportion of total number of samples represented by species i out of the total number of samples.
Ethical clearance was obtained from the Institutional Review Board (IRB) of Aklilu Lemma Institute of Pathobiology, Addis Ababa University, Ethiopia (Ref. No. ALIPB/IRB/012/2017/18) and National Ethics Review Committee (NERC), Ethiopia. Permission was also obtained from East Wollega and Buno Bedele Zonal Health Offices, Oromia Regional State, Ethiopia. Verbal consent was obtained from household owners to set CDC light traps.