Climate Change and Spatiotemporal Distributions of Vector-Borne Diseases in Nepal – A Systematic Synthesis of Literature

Background Despite its largely mountainous terrain for which this Himalayan country is a popular tourist destination, Nepal is now endemic for five major vector-borne diseases (VBDs), namely malaria, lymphatic filariasis, Japanese encephalitis, visceral leishmaniasis and dengue fever. There is increasing evidence about the impacts of climate change on VBDs especially in tropical highlands and temperate regions. Our aim is to explore whether the observed spatiotemporal distributions of VBDs in Nepal can be related to climate change. Methodology A systematic literature search was performed and summarized information on climate change and the spatiotemporal distribution of VBDs in Nepal from the published literature until December2014 following providing items for systematic review and meta-analysis (PRISMA) guidelines. Principal Findings We found 12 studies that analysed the trend of climatic data and are relevant for the study of VBDs, 38 studies that dealt with the spatial and temporal distribution of disease vectors and disease transmission. Among 38 studies, only eight studies assessed the association of VBDs with climatic variables. Our review highlights a pronounced warming in the mountains and an expansion of autochthonous cases of VBDs to non-endemic areas including mountain regions (i.e., at least 2,000 m above sea level). Furthermore, significant relationships between climatic variables and VBDs and their vectors are found in short-term studies. Conclusion Taking into account the weak health care systems and difficult geographic terrain of Nepal, increasing trade and movements of people, a lack of vector control interventions, observed relationships between climatic variables and VBDs and their vectors and the establishment of relevant disease vectors already at least 2,000 m above sea level, we conclude that climate change can intensify the risk of VBD epidemics in the mountain regions of Nepal if other non-climatic drivers of VBDs remain constant.


Methodology
A systematic literature search was performed and summarized information on climate change and the spatiotemporal distribution of VBDs in Nepal from the published literature until December2014 following providing items for systematic review and meta-analysis (PRISMA) guidelines.

Principal Findings
We found 12 studies that analysed the trend of climatic data and are relevant for the study of VBDs, 38 studies that dealt with the spatial and temporal distribution of disease vectors and disease transmission. Among 38 studies, only eight studies assessed the association of VBDs with climatic variables. Our review highlights a pronounced warming in the mountains and an expansion of autochthonous cases of VBDs to non-endemic areas including mountain regions (i.e., at least 2,000 m above sea level). Furthermore, significant relationships between climatic variables and VBDs and their vectors are found in short-term studies.
Although no single factor can fully explain the transmission of VBDs, climate change can alter the geographical distribution of disease vectors and VBDs [3], for example, by rendering previously endemic areas unsuitable and previously non-endemic areas suitable for their existence and reproduction [40,41,42]. As higher latitudes and altitudes are more sensitive to climate change and experiencing higher warming rates [2,43,44], a shift of species distributions towards higher elevations has been predicted [45]. Several recent studies have shown an increasing trend of the epidemic potential and length of the transmission season of VBDs in temperate regions and tropical highlands under different climate change scenarios [14,18,19,24]. Moreover, the IPCC concludes that local changes in temperature and rainfall will continue to alter the distribution of disease vectors and the risk of VBDs [4], and that the population mostly affected by climate-sensitive diseases and is deprived of participation in economic growth will face the largest health risk [4,46]. However, heterogeneity in VBD transmission is observed at every spatial scale ranging from less than a kilometre to continents. This heterogeneity is determined, for example, by the ecology and biogeography of the vectors, soil types, urbanization, local adaptation to temperature, and host communities [47,48]. Therefore, evidence generated at the global level may not be applicable at a local scale, and understanding the spatiotemporal distribution of VBDs in an individual country is important.
Nepal is one of the most vulnerable countries with respect to climate change because it is positioned in the southern rim of the so-called "Third Pole" of our planet (the Himalaya-Hindu Kush mountain range and the Tibetan Plateau). It has a complex topography and a low level of development [49]. Despite its largely mountainous terrain, Nepal is afflicted with five VBDs, namely malaria, lymphatic filariasis (LF), Japanese encephalitis (JE), visceral leishmaniasis (VL) (also known as kala-azar) and dengue fever (DF) [50]. Most of the major VBDs included in this study may be considered as neglected tropical diseases (NTDs). Their characteristics in Nepal are provided in Table 1.
The Nepal National Adaptation Programme of Action (NAPA) to climate change has identified VBDs as one of the highest priority adaptation projects forthe health sector in Nepal  [52,60] [49]. The higher warming trend in the hill and mountain regions of Nepal and increasing numbers of autochthonous cases of VBDs reported from non-endemic hilly and mountain regions motivated us to conduct this review. By reviewing the published literature we wished to explore whether these observed spatiotemporal distributions of VBDs can be attributed to climate change. Thus, this article aimed to review the available literature on VBDs and climate change to allow for an assessment of the likely impacts of climate change on the changing spatiotemporal distribution of VBDs in Nepal.

Methods
A systematic literature search was performed and summarized information on climate change and the spatiotemporal distribution of VBDs in Nepal following providing items for systematic review and meta-analysis(PRISMA) guidelines. We searched for peer-reviewed articles published in English language before December2014 in the PubMed and Web of Science databases. Besides, we searched for relevant journal articles in Google Scholar and retrieved government reports from their web sites. We used the following search terms in title, abstract and keywords: 1. Nepal and 2. Climate, climate change, temperature, rainfall, precipitation, relative humidity, weather, Aedes, Anopheles, Culex, Phlebotomus, dengue, malaria, kala-azar, Japanese encephalitis, visceral leishmaniasis, lymphatic filariasis, mosquito-borne diseases, VBDs.
We first screened "titles", "abstracts" and "keywords" for relevant articles and then read full text articles to evaluate them according to our inclusion criteria. Furthermore, the reference lists of each selected research article were then evaluated using the snowball sampling technique if they had been missed in the electronic databases. Inclusion criteria for selecting studies are listed below:

Inclusion criteria
1. Studies must include climatic variables (Rainfall, temperature and humidity) that analyse the trend of climatic data and are relevant for the study of VBDs 2. Epidemiological studies dealing with the spatial and temporal distribution of disease vectors and disease transmission and/or epidemiological studies assessing the association of VBDs with climatic variables 3. Only studies published before December2014 and with a study area in Nepal The selected papers were systematically reviewed and thematically analysed. Conference proceedings, viewpoint articles, review articles, project reports and theses were excluded. Given the low number of studies meeting the inclusion criteria and their mostly descriptive nature, a quantitative meta-analysis was not appropriate. Therefore, we alternatively summarized the state of the art on climate change and the spatiotemporal distribution of VBDs in Nepal. The spatio-temporal distributions of VBDs from the published literature was projected onto a map of Nepal using GIS software.
We found 12 studies that analysed the trend of climatic data and are relevant for the study of VBDs, 38 studies that dealt with the spatial and temporal distribution of disease vectors and disease transmission. Among 38 studies, only eight studies associate climatic factors with VBDs in Nepal. Among these eight studies, one dealt with malaria and VL, one with dengue and LF, three dealt with malaria and one each with JE, LF and VL.  Climate Change and Spread of Vector-Borne Diseases in Nepal A summary of the main findings from the analysis of climatic variables is provided in Table 3. Similarly, a summary of the main findings of studies on the spatiotemporal distribution of VBDs and their association with climatic variables is presented in

Observed climate change in Nepal
Although analyses of observed temperature and precipitation data are still limited in Nepal, climate change effects are already occuring. Temperature data show a warming trend in Nepal. This warming trend is influenced by maximum temperatures with higher warming rates in the mountain regions compared to the lowlands of Nepal [61,66,71]. A recent study using the data of 13 mountain stations of Nepal  reported that only the maximum and mean temperatures are in an increasing trend without any changing trend of minimum temperatures [69]. In contrast, another study conducted in the Mustang district of Nepal's mountain region using data between 1987-2009 shows an increasing trend of both minimum and maximum temperatures [72]. A general warming trend of all types of temperatures (minimum, maximum and mean) is observed in Nepal with a higher warming rate in mountain and hill regions compared to lowland (Terai) regions when the data of eight stations between 1971-2006 are analysed [64]. The annual cycle of temperature lapse rate shows a bi-modal pattern with two maxima in the pre-and post-monsoon seasons as well as two minima in winter and summer. This is completely different from other mountain regions suggesting different contributing factors in individual seasons [67]. Hence, the trend of temperatures, especially of minimum temperatures, depends on the location of meteorological stations because of large micro-climatic [67]. Precipitation does not show a distinct trend in Nepal. The analysis of precipitation data from 78 stations all across Nepal  and sub-regional records  reveal great interannual and decadal variability in precipitation but do not show a long-term trend [62]. The analysis of data from 26 stations (1961-2006) shows an increasing trend of extreme precipitation events in total and heavy precipitation at most stations [64]. A strong relationship between rainfall and elevation in the pre-monsoon and post-monsoon is observed along the Himalayas of Nepal [68]. Interestingly, two significant rainfall peaks are found over the  Central Himalayas (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008) Rainfall Spatiotemporal variation analysis based on satellite data A strong positive relationship between elevation and rainfall during the pre-monsoon season and two significant rainfall peaks during the summer monsoon season were found over the southern slope of the Himalayas; the primary peak along the Sub-Himalayas (~500-700 m asl) and the second peak along the Lesser Himalayas (~2,000-2,200 m asl) Only data from the pre-monsoon and summer monsoon seasons were included in analyses; the highest altitude considered was 5,000 m asl. The high resolution regional simulations were generated using the regional climate model PRECIS and were validated using observedAPHRODITE precipitation data and NCEP/NCAR temperature data.      southern slope of the Himalayas (between 500-700 and 2,000-2,200 m above sea level [asl]) during the monsoon season whereas relatively large amounts of rainfall occur over higher elevations during the pre-monsoon season [68]. Spatial and temporal variation in precipitation pattern and significant roles of mountainous relief in yielding localized precipitation patterns is reported in Nepal [63]. However, changes in the extreme events, consistent with climate change effects, are more significnat in Nepal. A declining trend of cool days and inclining trend of warm days are observed in the higher altitudes of Nepal [64]. The combined effects of increased temperature and diminished snowfall followed by rapid shrinking of the majority of glaciers have already resulted in a reduction of water available for drinking and farming in the mountain regions of Nepal [66,72,101]. The precipitation extremes show an increasing trend in total and heavy precipitation albeit no systematic difference is observed in extreme precipitation trends beween the highlands and lowlands [64].
The Hadley Centre's high-resolution regional climate model PRECIS (Providing Regional Climates for Impact Studies) projects significant warming towards the end of the 21 st century and a decrease in monsoon precipitation over the Central Himalayan region (which includes Nepal) during the period 2011-2040 and an increase in seasonal rainfall during the period 2071-2098 compared to the baseline period   [70]. In contrast, other studies with regional climate models such as COSMO-CLM project over 70% decrease in monsoon rainfall in parts of northern India at the end of this century [65] and because of higher evapotranspiration rates with higher temperature, a decreasing amount in water availability has also been reported. The fifth assessment report of IPCC conclude that different climate models have varying degree of success in simulating past mean state andclimate variability compared to observed sations data and large uncertainities in climate change projecttions are reported as estimated in multi-model ensembles (especailly for precipitation) [1]

Spatiotemporal distributions of vector-borne diseases in Nepal
Malaria. In Nepal, a heterogeneous spatial distribution and fluctuating trend of malaria incidence has been reported with a higher incidence in southern districts bordering India [90]. The distribution of the disease, which was previously confined to the forest and forest fringe regions of the Terai lowlands and so-called Inner Terai valleys and hills (<1,200 m asl) in 38 districts [86], is now observed to extend to the hills and mountains (> 2,000 m) [87] and malaria is now endemic in 65 out of the 75 administrative districts [89,102]. Fourty four species of Anopheles mosquitoes have been recorded in Nepal and out of these, seven have been incriminated in malaria transmission (Anopheles minimus, An. fluviatilis, An. maculatus, An. dravidicus, An. pseudowillmori, An. willmori and An. annularis). They were observed at least 2,000 m asl [95] and cause seasonal malaria epidemics in Nepal, including in areas above 2,000 m [51,52]. During the last decade, the incidence of confirmed malaria cases has declined significantly in Nepal following the introduction of the free distribution of long-lasting insecticide impregnated bed-nets (LLINs) and artemisinin combination therapy (ACT) for the treatment of Plasmodium falciparum malaria [89]. However, the proportions of P. falciparum and imported malaria cases have increased considerably in comparison to the total number of confirmed malaria cases [89]. This implies the possibility of a gradual shift in the Plasmodium parasite population possibly due to the rising temperature trends. Moreover, indigenous cases of P. vivax malaria and P. falciparum infections have been reported from the hill and mountain regions of Nepal in later years. A positive relationship between rainfall and malaria cases with a certain time lag has been observed in Nepal [87]. A significantly positive correlation of malaria incidence with minimum as well as maximum temperatures and rainfall was found in a study conducted in the high-endemic malaria district of Jhapa [88]. However, non-climatic variables were not included in the analysis in that study, and the climatic variables assessed were not significant predictors of malaria incidence in time series analysis. Another recent study shows that a1°C increase in minimum and mean temperatures increased malaria incidence by 27% and 25%, respectively. The reduction in malaria incidence was 25% per one unit increase of LLINs [23]. The spatiotemporal distribution of malaria in Nepal (1978-2012) is presented in Fig 3 [50,86,90,103].
Dengue. The first reported case of dengue virus (DENV) infection in Nepal was a Japanese volunteer in 2004 [73], and the first local transmission of DENV in Nepal was confirmed at the beginning of an outbreak in 2006 (August-November) in lowland urban areas of 11 districts [58,104]. During this outbreak, the first record of the primary vectors of DENV, Aedesaegypti mosquitoes, and the presence of all four serotypes of the virus (DENV1-4) were reported [58]. Previously, no Ae. aegypti had been reported in Nepal, but the secondary vector of DENV, Ae. albopictus was known to have existed in the lowlands and hill regions including Kathmandu (the capital city of the country located above 1,300 m asl) as early as the 1950s [52,91]. Since the first DF outbreak in 2006, DENV and its vector Ae.aegypti have been rapidly expanding across the country including the densely populated Kathmandu valley and mountain regions of Nepal [57,59,74,75,94,95]. During an epidemic in 2010, 917 DF cases including five deaths were reported [50]. During this outbreak, DF cases started to be recorded at the beginning of August (monsoon season) in a hilly district (~360 m asl), rapidly spread in lowland Terai Climate Change and Spread of Vector-Borne Diseases in Nepal districts (90 m), appeared in the hill districts of Dhading (> 900 m) and Kathmandu (> 1,300 m) in October with a peak in November (post-monsoon season) and diminished in mid-December (winter season) [59]. The study of the 2010 epidemic also recorded Ae. aegypti in all affected areas and provided data suggesting that DENV isolated from Nepalese patients was phylogenetically close to Indian DENV pointing to an import of the virus from India. Significant effects of the climatic factors temperature, rainfall and relative humidity, physiographic region and month of collection on the abundance of adult Ae. aegypti were reported in a recent study from central Nepal [105]. The spatiotemporal distribution of DF cases in Nepal in the period 2006-2012 is shown in Fig 4 [58 ,59,75,104].
Visceral leishmaniasis (VL). In Nepal, VL cases were first recorded in 1980. At that time confirmed cases were confined to lowland Terai districts of eastern and central Nepal that border India's state of Bihar, followed by records from 13 endemic districts and an increasing trend of incidence until 2003 [54]. Despite a declining trend of VL incidence in Nepal after 2003, VL is now increasingly reported from districts classified as non-endemic amounting by 2009 to 47 out of the 75 districts of Nepal (albeit cases were not classified as indigenous or imported) [98]. Although the disease was previously assumed to be confined to rural households with damp earthen floors and especially to poor families, autochthonous VL cases have since 1997 also been reported among residents of the urban area of Dharan city with highly clustered distributions [56]. Disease transmission in Dharan was confirmed by PCR identification of both vector, the sand fly Phlebotomus argentipes, and parasite, Leishmania donovani, inside town [56]. Moreover, series of autochthonous VL cases are now being reported from new areas mostly in hill and mountain regions of Nepal which had previously been considered to be non-endemic for this disease [96,97,99,100]. A positive association of VL cases with temperature and rainfall has been observed with reports of disease outbreaks 2-3 months after heavy rainfall in Nepal [87]. The abundance of the vector P. argentipes has also been found to be positively correlated with the maximum temperature of the month of collection and negatively correlated with the precipitation of previous months in both Nepal and India [106]. The spatiotemporal distribution of VL cases in Nepal (1980-2011) is shown in Fig 5 [54,96,97,98,99,100].
Japanese encephalitis (JE). Infections with Japanese encephalitis virus (JEV) moved northward in India and began to be seen in Nepal in the late 1970s [76] when epidemics occurred in lowland districts bordering India in western (Rupendehi) and eastern Nepal (Morang) in 1978 [78]. The mosquito species Culex tritaeniorhynchushas been reported to be the principal vector of JEV in many parts of Asia including Nepal [52,81]. It was first recorded in Nepal in 1965 with a distribution ranging from the lowland to hill regions including Kathmandu valley [60]. Although most reported JE cases in Nepal were initially confined to 24 districts in the lowland Terai [78], JEV transmission is now established in hill and mountain districts of Nepal, including Kathmandu valley, which were previously considered nonendemic for this disease [77,79,80,81,82]. Moreover, there are reports of spatial cluster of JE incidence with a shift from the Terai lowlands to hill and mountain regions after 2005 [81]. The risk of JE was also associated with paddy field configuration at the landscape level [83]. A presence of mosquitoes in pig farms and their association with JE sero-positivity has been reported from four mountain districts of Nepal [82]. A significantly positive association of JE incidence with monthly temperature and the percentage of irrigated land, and a negative association with low precipitation has also been reported from Nepal [81]. The spatiotemporal distribution of JE cases in Nepal (1978-2012) is shown in Fig 6[77 ,79,80,81,82,107].
Lymphatic filariasis (LF). The mosquito species Culex. quinquefasciatus, the principal vector of Wuchereria bancrofti microfilaria in South Asia, was first recorded in Nepal in 1956 [91] and found to occur within the LF endemic zones of this country [50,52,53]. In the year 2001, LF was endemic in 33 out of 37 surveyed districts of Nepal. The majority of cases were confined to an altitudinal range between 500-700 m asl, however, with a substantial number of cases at altitudes between 900-1400 m asl [84]. A sentinel surveillance conducted in 2007 among 7,000 people residing in six districts of the lowland (Terai), hill and mountain regions of Nepal reported the highest microfilaria infection rate (2.0%) in the mountain district of Sidhupalanchowk [85], suggesting a shift of LF transmission to the mountain region of Nepal after the introduction of mass drug administration (MDA) programmes in lowland and hill districts which had started in Parsa district in 2003. By 2013, six rounds of MDA had been completed in 16 endemic districts and four, three, two and one round of MDA in 10 districts each, and a gradual expansion of MDA reached 5 endemic districts covering 74% of the total population at risk (N = 21,852,201) [50]. However, 61 out of 75 administrative districts have already been reported as being LF endemic, and Nepal plans to cover the remaining six endemic districts with MDA by 2014 and achieve <1% prevalence in all endemic districts by 2018 [50]. Previously, Cx. quinquefasciatus mosquitoes had been recorded in all endemic districts ranging from 90 to 1,800 m asl [102], and recent studies report the distribution of Cx. quinquefasciatusup to at least 2,100 m (the highest sampled altitude in that study) in the districts of Dhunche and Rasuwa which had previously been regarded as non-endemic for LF [94], and above 2,000 m in Nagarkot of Bhaktapur district [93]. Moreover, significant effects of the climatic factors temperature and relative humidity, physiographic region and month of

Discussion
The review of observed and future projections of climatic data show a conducive environment for the transmission of VBDs in Nepal, especially in the highlands (mountains) which had been assumed to be free from these diseases. Despite a decade-long armed conflict and political instability in Nepal, there has been a substantial decline in the incidence of all major VBDs except DF which has only emerged in Nepal since 2004. The presence of disease vectors and reports of series of autochthonous cases of VBDs in hill and mountain regions of Nepal that had previously been considered to be non-endemic suggests that the local transmission of VBDs might be favoured by rising temperatures. However, the transmission of VBDs among humans is more complex than mere temperature changes, and this fact has been extensively reviewed [9,28,33,34,39,108,109,110,111]. One may also hypothesize that improvements in diesease surveillance and health care services, land use changes, population growth, globalization in general, in particular international trade, tourism and travel, migration and other movement of people, the expansion of road networks and the shipment of goods, unplanned urbanization and improvements in livelihood and access to health care services, etc., could be responsible for an increased detection of VBD cases in new areas. However, the presence of vectors and local transmission of VBDs at altitudes above 2,000 m, which clearly stands against the conventional logic that high altitude regions are free of VBDs because of cold temperature, strongly suggests that global warming is playing a role in the observed transmission. Therefore, we discuss below climate change and the spatio-temporal distribution of VBDs in Nepal in comparison with the results of studies from other mountainous countries around the world.
A warming trend of annual mean temperatures is observed throughout the country indicating that climate change is already occuringin Nepal. However, large spatial and temporal variation in the trends of minimum and maximum temperatures is observed across different meteorological stations. The warming signal is clearer for maximum temperature with a more pronounced warming in the mountains compared to the lowlands ofNepal. This is in sharp contrast to the warming trend of the Tibetan Plateau where the minimum temperature is increasing at a faster rate than the maximum temeprature [112,113,114]. but consistent with the trend observed in the western Himalayas of India [115,116], indicating a role of the Indian monsoon in the regulation of temperature through complicated feedback. Increasing trends of both minimum and maximum temperatures with greater warming rates in higher elevations have also been reported from the Rocky Mountains in Colorado, USA [117].
Precipitation is one of the major climatic factors affecting transmission of VBDs. The absence of a distinct long-term trend in precipitation changes in Nepal despite increasing GHG and strongly increasing aerosol concentrations in the region (especially through the neighbouring countries India and China) might be explained by a moister but less intense monsoon circulation [65]. Nepal's precipitation is affected by two major air movements: the summer monsoon which originates from the Bay of Bengal in the east, and the winter western disturbances which affect mostly the western parts of the country and result in snowfall in the mountains. A large spatial variation in annual rainfall over Nepal, ranging from less than 150 mm to more than 5,000 mm, is observed and is largely associated with the South Asian monsoon [118]. About 80% of the annual precipitation occurs during the monsoon season (June-September) followed by the post-monsoon season (12.7%) [118]. A positive correlation has been shown between the all-Nepal precipitation and the Southern Oscillation Index (SOI) series suggesting a strong association between the El Niño Southern Oscillation (ENSO) and precipitation fluctuation in Nepal [62]. The all-Nepal precipitation records do not agree well with the all-India precipitation record but resembles that of the northern part of India. This means that the precipitation climatology of the Himalayan region and adjacent areas differs greatly from the southern part of the Indian subcontinent. As a result, aggregated precipitation data from all over India cannot provide a valid representation of the entire subcontinent [62]. Although the regional climate models show an increasing trend of temperatures and no distinct trend in precipitation amount, most regional climate models report too warm temperatures in the northern parts of India that are too high compared to the observations, and the amounts of precipitation and its spatial distribution differ significantly between the regional climate models [119].
The observed declining trend of cool days and increasing trend of warm days in the higer altitudes of Nepal [64] is consistent with the global trend [43,44]. However, a multi-country study carried out in South Asia showed only the extreme temperature indices of low altitudes and latitudes to be consistent with general warming whereas stations at higher altitudes and latitudes showed both positive and negative trends, suggesting that high-elevation sites might be more influenced by local environmental factors making projections of future developments difficult [120]. On the other hand, a study from Nepal suggests that the contribution of the effects of urbanization and local land use changes to the all-Nepal temperature change is minimal, indicating that global warming is influencing the warming trend in Nepal which is comparable in magnitude with regional and northern hemisphere temperature trends [61]. It has also been argued that the diminishing snow and glacier covers in the mountain regions of Nepal will change the surface albedo of the region, which in turn will increase surface temperature, thereby acting as a positive feedback mechanism resulting in higher warming rates of maximum temperatures in the mountains compared to the lowlands [61,121]. The decreasing trend of the observed maximum temperature in the winter season of the lowland Terai regions is believed to be due to the occurrence of cold waves and prolonged periods of fog which have become more prominent in the last decades [64,118].
A review of elevation-dependent warming and its possible causes in four high mountain regions-the Swiss Alps, the Colorado Rocky Mountains, the Tibetan Plateau/Himalayas, and the Tropical Andes-showed variation in the trends of extreme temperatures suggesting the need for a comprehensive study analysing the minimum and maximum temperatures separately for all mountain regions together to better understand elevation-based warming in mountains [122]. The highly varied topography over short distances and poor coverage of observational datasets especially in mountain regions render spatiotemporal climate change projections for Nepal difficult and any projections must be interpreted with caution [123]. Nevertheless, uncertainty in climate projections should not be a barrier for assessing vulnerability, impacts and adaptation options.
Several studies predict an increasing trend of the epidemic potential and the transmission season of malaria in temperate regions due to climate change [14,18,19,20,21]. The duration of the malaria transmission window in India is predicted to increase in the northern and western states, and shorten in southern states under different climate change scenarios [124]. Accordingly, the temperate region of Nepal is predicted to be at risk of malaria due to climate change because it is experiencing a much higher warming trend compared to the sub-tropical regions, and sporadic autochthonous cases of malaria have been reported from mountain regions previously considered free of malaria risk. Reports of autochthonous cases of malaria in the highlands may be due to an establishment of vectors above 2,000 m asl (i.e., the altitudinal range might have increased) [51,52], continuous import of malaria cases due to an increasing movement and migration of people [50,89]and may be favoured by rising temperatures because P. falciparum has a higher temperature requirement than P.vivax (sporogonic temperature threshold for P.vivax is generally assumed to be 14.5-16.5°C and for P. falciparum16.5-18°C [21]. As Nepal is preparing to move toward malaria elimination with the ambitious goal of achieving this by 2026, the surveillance of malaria in areas previously considered to be low or no-risk areas should be strengthened. The malaria control in mountain regions is much more difficult compared to the lowlands owing to geographical difficulty, a scattered human population, poor coverage of health services and the fact that the majority of malaria cases in these regions are caused by P. vixax which has a high relapse/re-infection rate in Nepal [125]. Although the first autochthonous case of DENV in Nepal was confirmed at the beginning ofthe first outbreak in 2006, the case of a DENV infection in a Japanese volunteer to Nepal in 2004 [73] suggests that DENV was already being transmitted in Nepal prior to 2006. The rapid expansion of both DENV and its primary vector Ae. aegypti across the country, including mountain regions, within a short period of time [57,59,73,94,104,126,127] may be attributed to an introduction of mosquito eggs in used tyres, transport of adult mosquitoes in motor vehicles, and increased domestic and peri-domestic breeding opportunities due to increasingly frequent and long-lasting water shortages creating a water storage culture in urban areas like the Kathmandu valley. It may be further enhanced by rising temperatures and an influx of DENV via infected people who travelled to dengue endemic countries, e.g., India and other Southeast Asian countries [59]. The rapid expansion of dengue vectors in Nepal after 2006 may have been further supported by the end of the decade-long armed conflict in Nepal (1996)(1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006). After the end of the armed conflict in 2006, rapid urbanization, road expansion, trade and business and mobility of people increased in Nepal, all of which might have driven the rapid expansion of DENV in Nepal. The suppression of DF outbreaks between 2006 and 2010 may have been associated with a rapid consumption of old tyres (the major breeding containers of dengue vectors) in the major urban areas mostly of the Terai lowlands which were routinely burnt during frequent strikes and protests. A geographical expansion of DENV in this decade, since 2004, also occurred in another Himalayan country, Bhutan [128]. The expansion of DENV vectors towards mountain region in Nepal is consistent with findings from the Eastern Himalayas [129] and highlands of Mexico [130]. First autochthonous cases of chikungunya virus (CHKV) infection (also transmitted by Ae. aegypti and Ae. albopictus) were reported in Nepal in 2013 [131] and a CHKV outbreak in Bhutan in 2012 [132]. Consistent with the expansion of DENV and CHKV in temperate regions of South-East Asia, expansions of the ranges of the disease vectors and an increasing number of autochthonous cases of DENV and CHKV infections have been reported in Europe where both climate change and non-climatic factors have been reported as contributing factors [133,134]. The low knowledge of people on DF prevention and control in Nepal [105,135] coupled with weak diagnostic facilities, a poor case reporting system which does not incorporate the private healthcare sector, a lack of routine vector and national DF surveillance, and poor multi-sector coordination for DF control [75,136] can intensify DENV expansion and epidemics in Nepal.
Resistance development to first-line drugs against VL and the inadequate implementation of vector control interventions have been reported as the major causes for this increasing trend of VL in Nepal [54,137,138]. Important risk factors that promote VL transmission are poverty, housing styles (i.e., mud walls and damp floors in houses), the presence of cattle, and peridomestic vegetation [139,140,141,142,143]. As only patients with disease symptoms are eligible for treatment and many asymptomatic cases hidden in the community, the treatment of cases had almost no effect, in contrast to vector control, suggesting the need of effective vector control interventions [144]. Although a declining trend of VL in Nepal has been observed after 2003, the disease is mostly reported from endemic districts only by public health institutions while an increasing trend of VL cases from non-endemic districts has caused worries about its control and elimination in Nepal. The positive correlation between the abundance of the VL vector P. argentipes and the maximum temperature of the month of collection in Nepal and India [106] as well as the positive association between VL cases and annual rainfall and temperature [87], indicate an effect of climatic factors on VL transmission in Nepal. Impacts of climatic variability on the occurrence and transmission of leishmaniasis have been reported in many studies [27,145,146,147]. Hence, the recently observed expansion of VL into new areas of the country may be facilitated by climate change along with other factors and constitutes an obstacle to achieving the VL elimination goal by 2015.
As the reported major environmental factors influencing JEV transmission in Asia including Nepal are temperature and precipitation [81,148,149,150,151], climate change along with the JE vaccination campaign which started in 2006 the southern districts in Terai [79,81,152,153,154] may have already affected the spatial and temporal distribution of JEV transmission in this country. The presence of the principal vector of JEV, Cx. tritaeniorhynchus, and JEV circulation in higher altitudes of Tibet have also been reported [155]; this is consistent with reports of JEV transmission in the mountain regions of Nepal [79,80]. However, clinical cases presenting with non-malarial febrile or encephalitic syndromes are reported in increasing numbers every year in Nepal but their etiology is not known in the absence of studies on viruses other than DENV and JEV. For example, the isolation of Kunjin viruses (Australasian subtypes of West Nile Virus [WNV]) from the sera of domestic animals from districts with a low prevalence of JEV has suggested that WNV, too, may circulate in Nepal [156] and recent study shows evidence of the continued spread of WNV in Nepal [157]. Recently, outbreaks of Nipah virus have been suspected in eastern Nepal but still await laboratory confirmation.
The establishment of Cx. quinquefasciatus mosquitoes already above 2,100 m asl [94,127], frequent movement of people between endemic and non-endemic areas, low acceptance of MDA due to severe adverse effects in some people, and its low coverage in urban areas [102] pose challenges for achieving LF elimination in Nepal by 2020. In the future, this disease is likely to expand in mostdistricts of Nepal if integrated vector-control measures and active disease surveillance are not implemented. Although studies for the Indian sub-continent were not found, model projections show that climate change and population growth are dominant factors for predicting the risk of infection and spread of LF on the African continent [30,158]. However, predicting the transmission risk of LF under different climate change scenarios is difficult owning to the chronic nature of this disease and its association with the standard of living of people and environmental sanitation.
Although socio-economic development, medical care and vector-control measures can outweigh the influence of climate change on VBDs in some areas [13,159,160]., the worst effects of climate change will occur in the poorest and most vulnerable regions least benefitted by economic growth [46,159,160]. Accordingly, with the improvement in economic status and overall health indicators in Nepal, the reported incidence of all VBDs in Nepal has declined but reports of confirmed autochthonous cases of VBDs from new areas including mountain regions that had previously been considered non-endemic is worrisome. Against a background of weak health care systems, difficult geographic terrain, lack of vector-control interventions in the highlands, and continuous influx of infected people from disease endemic areas, climate change can intensify the potential risk of VBD epidemics in the mountain regions of Nepal. However, our review is not conclusive for a causal relationship between climate change and VBDs and need further research to determine attribution to climate change. Hence, it calls for further research using long-term data records and controlling possible confounders in analyses. The findings of our review are consistent with similar reviews on climate change and VBDs from neighboring countries like India and China as well as others [9,10,161,162]. These studies also suggest that the available data are inadequate for conclusions on the impacts of climate change on VBDs because of complex relationships with non-climatic factors and inconsistent findings in different geographical regions. Thus, we propose the following specific future research priorities taking into account the combined effects of climatic and non-climatic variables on disease vectors and VBDs: • Entomological, virological and parasitological research in different transects along an altitudinal gradient across the country to determine the presence of vectors and their role in disease transmission • Development of VBD risk maps for Nepal based on entomological, virological and parasitological evidence and climatic as well as land use to better guide the allocation of limited resources to the most vulnerable groups • Social-ecological and socio-economic research to identify the adaptation needs in different ecological regions and settings and plan public health preparedness, taking into account ethnic, religious, cultural and gender differences

Conclusion
The studies reviewed here suggest that both the observed and projected climate are conducive for the transmission of VBDs in the mountain regions of Nepal which had previously been considered non-endemic for these diseases. The short-term data shows a clear association between climatic factors and VBDs, but it is complex and difficult to project long-term effects of climate change in the face of rapid environmental and socio-economic changes and attribution to climate change is not determined in the existing studies. Despite continuous efforts of the government to control them and their declining incidence over the last decade (except for DF), VBDs have over the years been expanding their geographical ranges especially in mountain regions of the country. This might be attributed to environmental changes, in particular climate change, along with socio-economic factors. However, the observed spatial expansion of VBDs in new areas, especially in cool margins of mountain regions, that is correlated with the observed warming climate does not necessarily show a causal relationship. As VBDs show a heterogeneous distribution and spatiotemporal variation in the trends of climatic variables across the country, well-designed long-term local studies are needed to determine attribution of climate change to the observed transmission and distribution of VBDs in new areas. Therefore, VBD monitoring, surveillance and research should be strengthened in areas where risk of VBD is not yet determinedand VBD control programmes are not yet focused. Moreover, tourists and returning migrant workers coming to Nepal from disease endemic regions (including the country's own lowlands) should be made aware about VBDs, their responsibility and potential role in spreading infections especially when travelling in mountain regions, and should be encouraged to engage in reasonable preventive and prophylactic measures including vaccination.