The availability of the small LASCAR temperature and humidity logger has enabled collections of data from microhabitats and based on hourly measurements throughout the 24-hour cycle. In 2018, a group from Chennai in India used similar loggers to assess microclimate variables of the ambient environment to study the extrinsic incubation period of Plasmodium vivax and P. falciparum in an urban setting[14]. These useful instruments enable more precise ecological measurements in the field as we have also shown here. In several recent publications on malaria epidemiology relative humidity is mentioned, but often the measurements have been underestimated and often quoted from average numbers [8] or based on meteorological data from a collecting station[8, 15]. Even recently our own work in Zambia attempted to measure local relative humidity around villages near the Zambezi River using satellite information; it was difficult to transfer dew point measurements by the Landsat satellite to local relative humidity in a meaningful manner [16]. Obviously more precise meteorological measurements are needed.
The study design used here is somewhat unique. We have taken a broad climatic view covering large tracts of land, most of it prepared for agriculture. Climate is not necessary local, but often expands over large swathes of territory forming an expanded ecosystem. The southern province of Zambia is typical of much of south-central Africa. The land is somewhat undulating, with about 1000-2000 m elevation, and from the Congo (DRC) south, through the continent there is a moderate rainy season from November to March-April, a cooler May-June, and a hot dry period from August to November. The rainy season encourages malaria transmission, but the hot dry period is still critical when conditions permit mosquito foraging at night [17]. At that time day temperatures can be very high, approaching 40C, but with cooler nights[11]. Anopheline mosquitos have adapted to this climate, and as shown here and in a detailed study in Sudan[11], they can survive and forage under these conditions.
We have matched malaria incidence per 100,000 person-months determined at local health centres with the various prevailing climate conditions over the three years of measurements we have shown that in the dry season humidity does play a role, even in the absence of rainfall. We found that humidity during the dry season continues to have a significant effect on malaria incidence in our model. This is not trivial but indicates that aerial moisture can permit local transmission of malaria.
In the dry season, it is important to focus on climate when aerial conditions permit the insects to forage and survive, and note from where does the moisture come? As there is little or no rain in the hot season in Zambia and elsewhere, the moisture must come from subterranean ground water, or from transpiration of the tree and its leaves. Most African trees are not deciduous, they do not drop leaves seasonally but usually maintain live leaves through most of the year which transpire through the seasons. Local persistent aerial moisture will depend on the trees, movement of air, wind and storms and stillness. This means that under still conditions the overall air cover may become moist over much of the ecosystem. The area we have measured and recorded covers an area of several thousand km2 serviced by 13 clinics, here we included only nine stations (Fig. 1). The measurements are global, not individual, and indicate transmission, albeit at a very low level, and this is associated with periods of local humidity.
Water and aerial humidity have been associated with malaria since Roman times in the pontine marshes and latterly in the early Jewish settlements in Israel [18]. In Sudan, [11] careful collection of overnight resting places, showed the presence of blood-fed Anopheles gambiae sl (likely An. arabiensis) collected over a nine-month period which included the hot dry season [19] [11] and aestivating An. arabiensis females were found to survive the dry season. A recent paper from Thailand compared found a positive correlation with malaria incidence and relative humidity over four of the years[15].
Research carried out in the Sahel and adjacent countries [20] on members of the An. gambiae complex have been studied throughout dry seasons to investigate diapause, dormancy, migration and effects on the mosquito populations [7, 15. 17, 19 20]. Population studies involving genetic markers in A. arabiensis in Senegal showed little change in the genetic profile across the area from dry to rain season and back [10]. The authors consensus was that local survivors were derived from a permanent population deme spread over large areas that fluctuates seasonally. Our work tends to support these results. The natural transpiration of trees and bush vegetation recognized in the Zambian study demonstrates that throughout the dry periods there are times when there is sufficient moisture in the air to support mosquito foraging. This humidity can assist in the survival of pockets of females which are present in the area and when the rains occur in November this enables the mosquito population to expand rapidly. This is likely to occur over much of the African continent and needs to be considered wherever malaria control is practiced.
Limitations
The work reported here covers three continuous cycles of the hot dry period; data collected from nine operational loggers set in live active growing trees in close proximity of houses where people sleep and more closely represents the ecological features of the resting mosquito. The data collected has suggested that local aerial humidity affects transmission. However, as with most field based studies there are limitations. These include missing data due to broken or displaced loggers. Additionally, circumstances caused data collection to commence late in the first year of data collection, thus reducing the period of data collection. Furthermore, some of the infections may have been missed since the diagnoses were made using the malaria rapid diagnostic test which is less sensitive than blood smear examination, and also asymptomatic malaria infections are common during the dry season. Finally, our study was in Zambia, south central Africa, this area does represent much of the Southern Continent, however it needs to be repeated in these and other ecosystems where malaria occurs.