This study was conducted in the four health district of Côte d’Ivoire. The health districts are namely Aboisso (5° 28’ N, 3° 12’ W) and Bloléquin (6° 34’ N, 8° 00’ W) Odienné (9° 30’ N, 7° 33’ W) and Ouangolodougou (9° 58' N, 5° 09’ W) (Figure 1).
The district of Aboisso is located in the primary rainforest of Côte d’Ivoire on southwestern border with Ghana. The climate is humid tropical type, characterized by abundant rainfall with an average annual height of about 1,500 mm of rain. The average annual temperatures are between 25 and 27 °C. This district covers an area of over 4,662 km² with a population size of 307,852 people and, thus, a density of 66 inhabitants per km². Within this district, Anopheles mosquitoes were collected in the villages of Affiénou (5° 25’ N; 2° 56’ W) and Appouesso (5° 57’ N; 3° 10’ W). Both villages are drained by numerous streams and lowlands, and have modern and traditional housing, a health centre, electricity and water supply. Coffee, cocoa, rubber and palm oil are the main cash crops while vegetable, taro and banana are the main food crops in the area.
The district of Bloléquin is situated in the dense forest zone in the west of Côte d’Ivoire on the border with Liberia. The population is estimated at 123,336 inhabitants. The climate is mountainous type, with annual average rainfall sometimes exceeding 2,000 mm per year and annual temperatures ranging from 15 to 33 °C. It covers an area of 2,962 km2 with a population density of approximately 41 inhabitants per km2. The district is irrigated by important tributaries coming from the Sassandra and Cavally rivers, favouring the establishment of various permanent water courses, puddles and small dams. The study sites in this district covered the villages of Zéaglo (6° 34’ N, 7° 47’ W) and Dépouta (6° 54’ N, 7° 57’ W). Only Zéaglo has a health centre, electricity, water supply and modern housing. The agricultural activity is mainly based on coffee, cocoa and rubber culture. Food crops are dominated by banana, cassava, maize and vegetables, and included flooded lowland rice paddy and rainfed rice cultivation.
The district of Odienné is in the savannah zone in north-west of Côte d’Ivoire and shares a border in its western part with the Republic of Guinea. It covers an area of 14 000 km² with a population of 193,364, giving a density 13.8 inhabitants per km². The climate is tropical sub-humid type with annual rainfall varying between 1,400 and 1,600 mm per year and annual temperatures ranging between 25.4 and 33 °C. The vegetation is essentially dominated by savannah type vegetation, with trees or shrubs with a grassy tendency. In some places there are patches of forest and also forest galleries along the watercourses. The district is fed by tributaries of the Sassandra river such as the Bagoué and Tienba rivers. Our surveys were conducted in Gbéléban (09° 36’ N, 08° 08’ W) and Niénésso (09° 21’ N, 07°36’ W). Gbéléban is bordered to the south by the Gbanala river and has modern housing, a health centre, electricity and water, while Niénésso has no health centre and is bordered to the east by an undeveloped lowland which provides watering for cattle. Most of the local inhabitants are farmers and their staple crops include cereals, tubers, cotton and cashew nuts.
The health district of Ouangolodougou is in the savannah zone in the north of Côte d’Ivoire. It is bordered in the northern part by Burkina Faso and covers an area of 5,380 km2, with an estimated population of 260,519 habitants, giving a density of 48.4 inhabitants per km2. The district is characterized by a Sudanese climate with a unimodal rainfall regimen from May to November. The annual rainfall varies from 1000 to 1400 mm, while the mean annual temperature ranges from 21 to 35 °C. The minimum temperatures can drop to 16 °C due to the Harmattan wind during December and January. The natural vegetation is mainly a mixture of savannah and open forest characterized by trees and shrubs that are approximately 8–15 m in height. The soil is highly conducive to agriculture and most of the local inhabitants are farmers with staple crops including rice, maize, and cotton. Rice is mainly cultivated during the rainy season in flooded soils. The study area included the villages of Broundougou (9° 59’ N; 05°09'95''W) and Satolo (10° 10’ N; 05°27’W). Broundougou has a health centre, electricity, water supply and modern houses, whilst Satolo has no modern infrastructure.
Before this study, we explored the local malaria prevalence from 2013 to 2015 recorded in the yearbooks in each district to select the study sites. The prevalence of malaria varied from 22.5% to 23.5% in Aboisso and was estimated to be between 16.8% and 19.5% in Ouangolodougou. The prevalence varied from 38.2% to 41.7% in Bloléquin and 41.2% to 42.9% in Odienné.
Entomological surveys were conducted in two villages in each of the four cross-border health districts of Côte d’Ivoire identified above. We conducted the first phase of entomological collections in July-December 2016 and a second phase of collections during July-December 2017 to capture seasonal variations in mosquito species diversity and abundance and malaria transmission.
Adult mosquito collections
Adult mosquitoes were sampled from July 2016 to December 2017 using window exit traps (WETs) and pyrethrum knock-down spray collections (PSCs). In each site, 15 WETs were installed on the windows of inhabited houses for two consecutive days per survey. Mosquitoes in the traps were collected every morning between 6 a.m. and 9 a.m. PSCs were performed early each morning from 6 a.m. and 8 a.m., before opening the windows, in ten rooms selected in different dwellings during two days per site per district. PSCs were performed in households that were different from those used for WET collections. In case of unavailability or refusal of participants, mosquitoes were collected from neighbouring houses.
Field mosquito processing
Anophelines were sorted from collected culicines using morphological identification keys  Anopheline species were then determined morphologically . We dissected the ovaries of the females of Anopheles vectors (An. gambiaes.l., An. funestus, and An. nili) and observed the degree of coiling of ovarian tracheoles to determine their parity status . All collected anopheline females were stored individually in Eppendorf tubes containing desiccant, labelled with the study site, point and date of collection, and stored at -20 °C for further molecular analysis in the laboratory at the Centre Suisse de Recherches Scientifiques en Côte d'Ivoire in Abidjan, Côte d’Ivoire.
Molecular identification of Anopheles gambiae complex members
DNA was extracted from the legs of An. gambiae mosquitoes using the boiling preparation methods . Briefly, three legs of each female Anopheles were cut and crushed in 100 mL of distilled water and boiled at 95°C for 10 min. The supernatant was pipetted and transferred into new labelled Eppendorf tubes which were stored at -20 °C and used as template for the polymerase chain reaction (PCR).
An. gambiae complex members were identified according to the SINE-PCR molecular method described by Santolamazza et al. . The primer F6.1A with the sequence 5′-TCGCCTTAGACCTTGCGTTA-3′ was used to distinguish An. coluzzii and the primer R6.1B with the sequence 5′-CGCTTCAAGAATTCGAGATAC-3′ to distinguish An. gambiae s.s. A LongGene® thermocycler (A200 Gradient Thermal cycler; LongGene Scientific Instruments Co., Ltd Hangzhou, P.R. China) was used with the following programme: 37 °C for 30 min, 94 °C for 30 s, and 59 °C for 30 s; 72 °C for 1 min repeated 35 times; and a final step at 72 °C for 10 min to finish the reaction. An agarose gel was prepared with 2% agarose in TBE (Tris/borate/EDTA) containing ethidium bromide at 10 mg/ml. The PCR product was loaded onto the agarose gel and allowed to migrate under a voltage of 100 V for 70 min. The result was visualized with a UV illuminator (TOYOBO Trans Modele TM-20).
Determination of sporozoite rates in Anopheles gambiae
DNA was extracted from the head and thorax of the adult females of Anopheles vector mosquitoes in each district  and screened for Plasmodium DNA using the fast COX-I PCR method described in Echeverry et al. . This is a very sensitive and rapid method which uses a set of primers, COX-IF (5′AGAACGAACGCTTTTAACGCCTG 3′) and COX-IR (3′ ACTTAATGGTGGAT ATAAAGTCCATCCwGT 5′), to amplify a polymorphic fragment of the COX-(I) gene using a recombinant DNA polymerase. The thermocycler (TaKaRa PCR Thermal Cycler Dice TP600) program was 94 °C for 5 min followed by 40 cycles of 94 °C for 1 min, 62 °C for 1 min and 72 °C for 90 s, and a final elongation step at 72 °C for 10 min. Five microlitres of the PCR product was visualized on 1% agarose gel in order to confirm amplification of the expected ~540 bp product (Plasmodium genus positive).
Malaria infection prevalence
In each health district, data on the malaria prevalence in the local populations were collected from the MoH statistical yearbooks. The total number of consultations was compared with the number of people examined with malaria from 2016 to 2017.
Data were double entered in Microsoft Excel 2013 software and transferred to STATA 14 (Stata Corp, College Station, Tx, USA) for analysis. The Kruskal–Wallis (KW) test was used to compare the differences in mosquito densities. The parity rate (PR) was calculated as the number of parous females multiplied by 100 and divided by the total number of females dissected. The P. falciparum sporozoite infection rate in each vector species population was calculated by dividing the number of Plasmodium-positive mosquitoes by the total number of mosquitoes tested, and this was expressed as a percentage (%). The χ2 test was used to compare sporozoite and PR between the collection sites and the health districts. All differences were considered significant at p < 0.05.