Effects of environmental temperature on parasite development (Extrinsic Incubation Period)
Using the thermodynamic parasite development model, we estimated the effect of using the mean daily temperature and mean hourly temperature on the calculation of the EIP for each site (1800-3200m). Our methods described the rate of parasite development for Plasmodium species (two human and one avian Plasmodium ), Haemoproteus , andLeucocytozoon parasites (Figure 3) at varying temperatures. For high elevation sites (3200m), the EIP estimates for all parasites using both measures suggested that transmission was not possible throughout the year. Below, we describe where results from both measures coincide, for other elevations and parasites, as well as where they differ in detail. We find that both measures, based on either the diurnal temperature range or the mean temperature, very largely give similar results for the periods of the year where the EIP is within the allowable range for transmission. However, for specific months, it is possible that transmission is possible within one scheme but not within the other (Figure 4).
For the 2600m site, there was no parasite transmission predicted using the EIP based on both temperature measures for P. falciparum andP. vivax from May to March. However, in June, the mean temperature range predicted no transmission of P. vivax whereas the EIP days of 46.9±15.71 using the diurnal temperature range. In April, the EIP days for P. vivax were longer by 13.95 days using the diurnal temperature range (EIP days: 36.08±3.5) than the mean temperature (EIP days: 22.13±4.04). The EIP days for P. falciparum were 5.62 days shorter using mean temperature measures (EIP days:40.29± 12.11) than the diurnal temperature (EIP days: 45.91±10.64) in April.
For avian Plasmodium relictum , using both mean temperature and diurnal temperature range, September to March was predicted as no transmission period and May to August and April as predicted transmission, at 2600m. In April, the EIP was 18.7 days longer using the diurnal temperature range (EIP days:32.95±16.29) than the mean temperature range (EIP days:14.25±2.25). For two malaria-like avian parasites, Haemoproteus and Leucocytozoon , September to March were predicted as no transmission period and June to August was predicted as the transmission period (< 1 day difference in EIP estimates) using both temperature measures. However, diurnal temperature range predicted no parasite transmission in April whereas mean temperature range estimated the EIP days of 8.27±1.1 and 5.62±0.75, for Haemoproteus and Leucocytozoon , respectively.
For the 2000m site (Anusuya), October to March was predicted to have no transmission period for P. falciparum and P. vivax using both temperature measures. In April, using the mean temperature predicted EIP days of 37.12±27.34 for P. falciparum and 34.73±9.66 for P. vivax , however no transmission was obtained using the diurnal temperature range. Using the mean temperature, there was no P. falciparum transmission predicted for May and September. However, the diurnal temperature range approach predicted EIP days of 51.7±17.94 in May and 50.7±17.02 in September.
Among avian parasites, P. relictum transmission window was predicted from May to September using both temperature measures. However, there was no P. relictum transmission predicted in October using mean temperatures while the diurnal temperature range method estimated EIP days of (47.4±19.46). In March, use of the diurnal temperature range predicted no P. relictum transmission whereas the mean temperature range measure suggested EIP days of 46.03±17.43. In April, both temperature measures predicted P. relictumtransmission, however, EIP days using diurnal temperature range were 3.06 longer than for the mean temperature. For two malaria-like parasites, Haemoproteus and Leucocytozoon, transmission was predicted from June to September using both temperature measures. There was no Haemoproteus transmission predicted for October to April using both temperature measures. In May, Haemoproteusshowed no transmission using mean temperature whereas, using the diurnal temperature, transmission was predicted with showed EIP days of 15.44±8.54. However, Leucocytozoon transmission was not predicted from November to February, while only the mean temperature supported EIP days of 17.71±13.56 in March and April. The diurnal temperature range method indicated transmission in October with EIP days of 19.81±7.8.
For the lowest elevation site (1800m), P. falciparum and P. vivax showed no transmission from December to February using both measures, the DTR and the mean temperature. An exception was for P.vivax where the use of the diurnal temperature range showed an EIP of 51.6±10.72 days in the month of February. The effect of daily temperature variations on P. falciparum and P. vivaxpredicted a transmission window from May to October, March, and April. In November, there was no transmission using the mean temperature for both P. falciparum and P. vivax. The use of the diurnal temperature range predicted transmission for both parasites in that month, with EIP more than 45 days in both cases.
In contrast, for P. relictum , use of the diurnal temperature range suggested transmission throughout the year. However, the use of the mean temperature showed no transmission across the months December to February. For the two malaria-like parasites, transmission months ranged from May to October for Haemoproteus and May to November, as well as April, for Leucocytozoon. The exceptions were: using the mean temperature led to transmission in March while using the diurnal temperature range led to transmission in April forHaemoproteus . In March, there was no transmission predicted using the diurnal temperature range while mean temperature range predicted EIP days of 12.79±11.96 for Leucocytozoon .
Our comparisons of mean temperature collected using experimental logger data and WorldClim data from 2014-2015 showed threshold temperature not supporting parasite transmission at the high elevation site (3200m) throughout the year (Suppl. Fig. S1). However, at the 2600 m site,P. vivax transmission was predicted from May to September and April whereas experimental data suggested only in April. For avian parasites, P. relictum , Haemoproteus andLeucocytozoon the window in which transmission is predicted is largely similar using these two source datasets, albeit leading to longer EIP days with experimental data, with the only exception for the month of April where this systematics are reversed. This larger pattern was reversed at 1800 m, with shorter EIP days with experimental data as compared to using WorldClim data, with the only exception again being for the month of April. The 2000m site showed closely similar transmission patterns for all parasites using both datasets, again except for the months of March and April, for P. relictum andLeucocytozoon .