Climate change and waterborne diseases in temperate regions: a systematic review

Recent global changes in climatic conditions have resulted in diverse changes in temperature and precipitation patterns in different regions of the world. For instance, heavy precipitation has increased in most parts of North America, most of Europe, much of southern South America, most of the Indian subcontinent, parts of northern and southeast Asia, parts of Australia, and parts of southern Africa since the 1950s (IPCC 2012; Dunn et al. 2020; Caretta et al. 2022). On the other hand, there has been a decline in intense precipitation events in eastern Australia, western Africa and northeast South America, which are mostly tropical regions (IPCC 2012; Dunn et al. 2020; Caretta et al. 2022). Within each region, there are also marked variations such as increases in annual precipitation in northern and eastern parts of North America as opposed to a decline across the western parts of the region in the last few decades (IPCC 2022). Furthermore, temperature increases within the North American region over the past few decades have resulted in a decline in snowpack and snow extent across much of Canada and the western United States of America (IPCC 2022). Mean temperature increases have also been measured across Europe, especially southern Europe, since the 1950s (IPCC 2022).

Reports indicate that approximately 3.6 billion people currently reside in areas that are highly susceptible to climate change; and about 250,000 additional deaths are projected to occur annually from the burden of climate-induced diseases between 2030 and 2050 (WHO 2023). Risks of waterborne diseases (WBDs) are likely to be influenced by changes in the quantity and the quality of freshwater resources as a result of changes in climate (IPCC 2014; Levy et al. 2018). An increase in frequency and intensity of extreme weather events such as floods and droughts may lead to an increased risk of foodborne diseases (Brubacher et al. 2020). Increasing temperatures may result in potential risks through emerging or re-emerging vector borne diseases (IPCC 2014; Trájer et al. 2022). Increased risks of respiratory tract infections may arise through temperature and humidity extremes, dust storms and extreme precipitation events (Cunsolo Willox et al. 2012; IPCC 2022; Trájer et al. 2022). Risk associated with vector borne, zoonotic and other infectious diseases may also be prone to climatic-induced land use and habitat changes (Cunsolo Willox et al. 2012; Sanchez et al. 2021). Furthermore, reports on non-communicable diseases exhibiting climate-sensitive health risks are documented, including risks of mental and emotional health issues as a result of climate-induced displacement (Cunsolo Willox et al. 2012; Bertone et al. 2016; WHO 2023); risk of physical injuries due to heat and extreme weather events such as flooding and fires; and risk of cardiovascular diseases through high temperatures and extreme heat (Cunsolo Willox et al. 2012; IPCC 2022).

Common WBDs documented in the literature include cholera, typhoid, diarrhoea, giardiasis, cryptosporidiosis and leptospirosis (IPCC 2022). Infection occurs when drinking water containing infectious doses of pathogenic microorganisms (bacteria, protozoa, viruses and parasites) (IPCC 2022). Although there are variations by country, the global burden of WBDs has declined due partly to improved sanitation and hygiene. However, risk of WBDs still persists and poses health challenges even in temperate regions (Levy et al. 2018; Brubacher et al. 2020). For instance, an estimated seven million waterborne-related illnesses were reported annually in the United States between 2000 and 2015 (Collier et al. 2021), while in 2015, campylobacteriosis, salmonellosis and infection with Shiga toxin-producing Escherichia coli (STEC) represented more than 75% of the burden of foodborne and waterborne diseases reported in Europe (Cassini et al. 2016).

Complex relationships between weather conditions and health risks for humans call for a better understanding of the influence of climatic factors such as rainfall and temperature on WBDs. A global review documented the effects of extreme precipitation or temperature on drinking water-related waterborne infections from 2001 to 2013, and most studies identified a positive association between increased precipitation or temperature and waterborne outbreaks or cases (Guzman Herrador et al. 2015; Levy et al. 2016; Philipsborn et al. 2016). A similar study that examined links between extreme water-related weather events and WBDs between 1910 and 2010 characterized heavy rainfall and flooding as important attributes preceding outbreaks (Cann et al. 2013). It further identified the most prevalent waterborne pathogens globally as Vibrio spp. and Leptospira spp. However, what is unclear is the extent of influence of climatic factors on the risk of specific WBD in a changing climate in temperate regions (Forbes et al. 2021). In recent decades, climate-change impacts have become more frequent and intense in temperate regions (IPCC 2022). However, knowledge of dynamics of climate-pathogen relationships and the associated health risks posed by specific WBDs is still limited in these regions (Galway et al. 2015; IPCC 2022). Against this backdrop, this study seeks to identify how climate change is influencing existing relationships between climate-related factors and specific waterborne pathogens/diseases in temperate regions in the literature; identify projected rates of waterborne infections through future climatic scenarios; identify methods or models used to project future WBD risks in a future climate in temperate regions; and highlight emerging themes observed from the review.

The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement was used as a guide for this review (Moher et al. 2009; Page et al. 2021). Based on the purpose of this study, keywords (Table 1) were identified as search terms reflecting freshwater sources, causative pathogens of diarrhoeal diseases, climate-related drivers and the scope of region. Scopus and Web of Science abstract databases were used given their broad inclusion not only of biomedical abstracts, but multidisciplinary research domains associated with climate drivers and model development. Temperate regions are defined as countries located in temperate climatic zones spanning North America, Europe, parts of Asia, and Australia/Oceania (see Table 1). Peer-reviewed full-text articles in the English language, published between 2010 and 2023 were included in the study. The chosen time frame aligns with a period of landmarks in tackling climate change, such as the Cancun Agreements of 2010 and the 2015 Paris Agreement on Climate Change (UN 2018). Exclusion criteria included non-freshwater sources, non-English language publications and locations outside of temperate regions, i.e., humid subtropical climate zones (such as Japan and Korea) and countries with a mix of climatic zones (such as Argentina and Iran). Study limitations include the restriction to English language publications which was associated with lack of resources for translation. It is acknowledged that this may have excluded some potentially useful publications and may result in lack of inclusion of research from some countries in temperate climate regions.

Data from the 75 included articles were extracted using a predefined template in Excel^©. The template included bibliometric information, region of study, disease or pathogen, climate-related factors considered, methods, model types, climate scenario information, disease or pathogen predictions as well as other useful information (Supplementary material, Table S1). A manual data analysis was conducted using a thematic analysis (Thomas & Harden 2008) to identify important themes that answer the question on how climate change impacts WBDs in temperate regions. Relevant data from the included studies were coded into the various themes in the database. For quality control, two independent reviewers each assessed a sample of three similar full-text articles from the 75 included articles as a pilot test to ascertain the quality of the data extraction by extracting data from the three articles using the predefined matrix. The two independent reviewers then came together to compare their data extractions and both reviewers were satisfied with the high semblance of both reviewers' views. The overall process was checked by a third reviewer.

The most studied bacteria (and overall pathogen) was E. coli, while the most prevalent protozoa were Cryptosporidium and Giardia (see Table 2). Viral pathogens studied were either recorded once or twice as secondary pathogens. Conversely, cryptosporidiosis was the most reported waterborne disease, followed closely by giardiasis, and gastrointestinal and waterborne illnesses/infections (see Table 3).

Identified sources of pathogens included large animals/wildlife (Cunsolo Willox et al. 2012; Eregno et al. 2016; Invik et al. 2019), waterfowl (Marcheggiani et al. 2010; Allende et al. 2017), droppings of animals (Tryland et al. 2011; Mohammed et al. 2019b), agricultural activities (Cunsolo Willox et al. 2012; Paruch et al. 2015; Coffey et al. 2016; Corsi et al. 2016; Allende et al. 2017; Buckerfield et al. 2019), forested catchments (Harris et al. 2018; Mohammed et al. 2019b), multiple land use (Galway et al. 2015; Eregno et al. 2016; Brunn et al. 2019), leakages and seepages from households' septic systems (Tryland et al. 2011; Mohammed et al. 2019b), combined sewer overflows (Eregno et al. 2018), wastewater and sewage treatment plants (Aguirre et al. 2016; Corsi et al. 2016; Bojarczuk et al. 2018), and human faecal contamination of separated stormwater systems (Petterson et al. 2016).

There are observed and projected changes in intensities and frequencies of precipitation events in temperate regions due to changes in climate. This is elucidated in direct and indirect effects of precipitation events on WBD risk.

Rainfall and heavy rainfall events accounted for the most frequently identified climatic driver of waterborne pathogens and diseases observed in this review (Figure 4). E. coli variations and peaks were found to be primarily linked to cumulative rainfall or changes in rainfall (Bougeard et al. 2011; Henry et al. 2016; Allende et al. 2017; Boudou et al. 2021). Many studies showed increasing loads of E. coli in water sources with increasing precipitation, especially during and after intense and heavy precipitation events (Tryland et al. 2011; Amirat et al. 2012; McCarthy et al. 2013; Jalliffier-Verne et al. 2015; Paruch et al. 2015; O'Dwyer et al. 2016; Harris et al. 2018; Herrig et al. 2019; Herrador et al. 2021). One study that showed an increase in loads of E. coli immediately after heavy rainfall events also observed reductions to prior baseline levels within 7 days in rainwater storage systems (Martin et al. 2010) while another observed variations in stormwater (Pieniaszek et al. 2021). Low-intensity rainfall in the summer was also found to increase E. coli in small streams by an order of magnitude (Oliver et al. 2015).

Parasitic protozoa (Cryptosporidium and Giardia) were only detected in surface waters during or after rain events (Tryland et al. 2011; Paruch et al. 2015). Heavy rainfalls have been associated with increased Cryptosporidium concentrations in water reservoirs (Bertone et al. 2016); increased risks of cryptosporidiosis (Boudou et al. 2021), and of cryptosporidiosis and giardiasis (Britton et al. 2010). Extreme precipitation as an annual rainfall anomaly was also associated with an increase in cryptosporidiosis risk in Australia (Forbes et al. 2021). Likewise, there was an increased risk of cryptosporidiosis and giardiasis cases when extreme precipitation was preceded by a dry period as observed in the Great Lakes region of North America (Graydon et al. 2022) and drinking water systems in Vancouver, Canada, with a lag of 4–6 weeks (Chhetri et al. 2017, 2019).

This suggests that the intensity of such precipitation events creates favourable conditions for the survival and transport of waterborne pathogens through increased runoff into water sources. As zoonoses, Cryptosporidium and Giardia contamination will result from excretion of animals being washed into the water source through overland and shallow groundwater flow due to the intensity of extreme precipitation (Tryland et al. 2011).

In rural areas, an increase in E. coli and total coliforms in untreated wells was influenced by an increase in precipitation along with agricultural drivers such as increased zero-till farming, increased sand content, and higher density of large animals (Ratajczak et al. 2010; Invik et al. 2019). These studies highlight the potential for increased contamination of groundwater when heavy rainfalls transport infected excreta into the subsurface, potentially increasing health risks in populations that use water from these sources. Similarly, Salmonella concentrations in irrigation ponds were higher after rain events in rural settings due to increased storm runoff from forested catchments (Harris et al. 2018). Rainfall–runoff events were also critical drivers of high concentrations of waterborne pathogens in stormwater retention basins within a German (Schreiber et al. 2019) and an Australian (Sidhu et al. 2012) catchment; and from snowmelt events (Galway et al. 2014).

Indirect effects were observed through an increase in spatial and temporal variations of E. coli concentrations in bivalves (mussels and Pacific oysters) in freshwater estuaries and rivers, respectively, significantly influenced by increasing rainfall levels with a lag time of 3–6 days prior to the sampling date (Campos et al. 2011; Colaiuda et al. 2021), and in shellfish fishery waters (Gentili et al. 2014). Similarly, a significant increase in Cryptosporidium oocysts count was observed in commercial shellfish harvest zones in a river in Prince Edward Island, Canada, after an intense rain event that was almost two times higher than after a low-intensity rain event (Aguirre et al. 2016). In simulated scenarios, rainfall was identified as a strong predictor of E. coli levels at harvest of baby spinach in irrigated waters whereas in scenarios where rainfall was excluded in the winter and the spring, E. coli counts and percentage of positive samples at harvest significantly declined (Allende et al. 2017). This may suggest the influence of water input as a significant driver, either in the form of natural rainfall or irrigation water, on increasing the potential of pathogen transfer from the organic matter or manure in the soil to the plant. The interplay between rainfall and irrigation is important within the context of climate change as irrigation is likely to increase in regions projected to become drier in the future, while rainfall is projected to increase in other regions.

High-risk levels of viral pathogens including human and bovine viruses, norovirus and enteroviruses are also recorded after heavy rainfalls in freshwater beaches, suggesting the influence of multiple contamination sources in their catchments and distinct beach environments (Corsi et al. 2016; Eregno et al. 2018, 2016; He et al. 2019; Sanchez et al. 2021). Distance to stormwater outlets also influences bacteria concentrations in bathing waters as Hong et al. (2021) observed higher E. coli concentrations in a recreational lake at the points closer to the stormwater outlet versus lower concentrations at farther points. However, rainfall was negatively associated with E. coli levels in Niagara beaches in contrast with the beaches in Toronto (Sanchez et al. 2021).

Predicted increasing rainfall intensity suggests a potential increase in waterborne infections especially in catchments with elevated agricultural land use such as crops, animal farms, or poultry (Tryland et al. 2011), and in small drinking water systems (Paruch et al. 2015; Wang et al. 2018; Invik et al. 2019). After days of extreme precipitation or heavy rainfall, hospitalization risks of waterborne enteric diseases increased by an hazard ratio of 1.73 among young children (Lai et al. 2020) and increased in populations within catchments of vulnerable karstic water sources (Dura et al. 2010).

An association was observed between infectious gastrointestinal illnesses (IGIs) and large volumes of water in the spring and summer seasons in the high north of Canada (Harper et al. 2011). Peak IGI-related clinic visits were preceded by high levels of water volume input 2 and 4 weeks prior in rural areas that depend on raw surface water for their domestic needs (Harper et al. 2011). Conversely, heavy precipitation events were associated with a decreased risk of gastroenteritis outbreaks (Herrador et al. 2021), likely due to dilution effects.

Wet periods or their onset played significant roles in E. coli concentrations (Ratajczak et al. 2010; Paruch et al. 2015; Buckerfield et al. 2019) and in Cryptosporidium and Giardia concentrations in surface water (Paruch et al. 2015), and in Cryptosporidium in wastewater treatment plants (WWTPs) (Xiao et al. 2022). Conversely, periods of dry weather had significant influence on E. coli concentrations in lagoon water (Derolez et al. 2013) and on faecal indicator bacteria (FIB) which exceeded managed limits of recreational waters (Sidhu et al. 2012); while viral pathogens (human adenovirus and polyomavirus) were prevalent in recreational waters during both wet and dry weather (Sidhu et al. 2012).

Temperate regions are noted for four distinct seasons each year. Differentiated seasonal risks have been identified in the literature to date. The summer season played significant roles in the risk of common gastrointestinal illnesses (David et al. 2014; Galway et al. 2014), and in IGI-related clinic visits in rural communities which rely largely on raw surface water (Harper et al. 2011). In the same vein, increased levels of E. coli contamination in shellfish waters (Pereira et al. 2015), and increased risk of infectious diarrhoea after a severe flood event in 2016 in China (Zhang et al. 2019) were identified during summer. Late summer is also associated with a peak in incidence rates of giardiasis (Brunn et al. 2019). In the spring season, increased levels of E. coli contamination in shellfish waters were also distinctly identified (Pereira et al. 2015); and higher infection rates of Cryptosporidium was more prevalent in the spring season than in summer and autumn (Xiao et al. 2022). However, spring conditions had less influence on E. coli levels in irrigation waters than during winter (Allende et al. 2017).

Autumn season, or its peak, coincided with an increased infectious diarrhoea risk after a severe flood event in 2016 in China (Zhang et al. 2019). This suggests an association between infectious diarrhoea risk and summer or autumn season, especially after severe flood events. Increased risk of acute gastrointestinal illnesses (AGIs) (Galway et al. 2014) and increased pollution and contamination of surface water with E. coli (Amirat et al. 2012) were observed in autumn. Models predict an approximately three-fold increase in average E. coli concentrations by 2075 for the spring and autumn seasons than current levels (Mohammed et al. 2019a). Higher infection intensity of Cryptosporidium oocysts in WWTP was observed during winter (Xiao et al. 2022).

Associations between streamflow and WBD risk have been documented in the context of observed and projected changes to streamflow. Streamflow was the most often identified hydrological variable associated with WBDs (Figure 5). It had significant influence on the risk of E. coli (Campos et al. 2011; Schernewski et al. 2012; Jalliffier-Verne et al. 2017, 2015; Bojarczuk et al. 2018; Mohammed et al. 2019a; Colaiuda et al. 2021), total coliforms (Bojarczuk et al. 2018), Cryptosporidium (Swaffer et al. 2014; Galway et al. 2015), cases of giardiasis (Brunn et al. 2019), Giardia (Galway et al. 2015), Campylobacter (Galway et al. 2015), bacillary dysentery in specific hilly/mountainous regions of Anhui province in China (Zuo et al. 2021), and transport or spread of FIB (Henry et al. 2016; Herrig et al. 2019). Specifically, when mean monthly streamflow increased by 1 m³/s, there was a 15% increase in AGI cases in the same month (Galway et al. 2015).

Other variables identified as associated with WBD risk include water levels/depth (Harper et al. 2011; Cunsolo Willox et al. 2012; Francavilla et al. 2012; Bojarczuk et al. 2018), water temperature (Francavilla et al. 2012; Masina et al. 2019; Mohammed et al. 2019a; Sanchez et al. 2021; Sokolova et al. 2022), turbidity (Swaffer et al. 2014; Sanchez et al. 2021), wave direction/height (Corsi et al. 2016; Sanchez et al. 2021) and unconsolidated aquifer (Brubacher et al. 2020).

There is some evidence that floods are associated with WBD risk. This is important to elucidate given the changes in flood frequencies and intensities observed and predicted for some regions due to changes in climate. For instance, in Ireland, a link was observed between peaks of E. coli enteritis and cryptosporidiosis cases (in April, 2016) and the 2015/16 winter flood event which began approximately 18 weeks earlier (Boudou et al. 2021). Likewise, at a national level in Italy, an association was identified between flood events and a medium-category of WBD risk effects when a retrospective study of 24 flood events (1993–2003) and registered cases of WBDs and their locations were analyzed (Marcheggiani et al. 2010). Also, the severe flood in China in 2016 had a significant influence on the increased risk of infectious diarrhoea risk predominantly among populations exposed to the flood events (Zhang et al. 2019). Associations between flood events and WBD risk were also observed with significant increases in the risk of Cryptosporidium in Germany (Gertler et al. 2015), and bacillary dysentery in China (Zhang et al. 2016). Furthermore, a link was observed between populations exposed to sampled urban floodwaters (within a 12-month period) and WBD risk in the Netherlands (de Man et al. 2014).

Given projections that temperate regions will become warmer because of a changing climate, it is important to shed light on associations between temperatures and WBD risk. Low water temperatures (1.1–6.7 °C) and low air temperatures (−0.1 to 4.5 °C) were significantly linked with an increased odds ratio of 0.047 and 0.041, respectively of Giardia and Cryptosporidium presence in untreated drinking water sources in Nunavut, Canada (Masina et al. 2019) and with increased concentrations of Cryptosporidium oocysts in WWTP (Xiao et al. 2022). Water temperatures were also shown to be essential predictors of E. coli in surface waters used for drinking water supplies as E. coli concentrations were higher in the colder periods than in the warmer times, which may have resulted from higher runoff intensity and subsequent increase in transport of pathogens into the catchment (Sokolova et al. 2022) and at the intake point of a lake supplying the water treatment plant in a future scenario (Mohammed et al. 2019a).

High water temperatures (8–29 °C) had a negative influence on the survival of Giardia and Cryptosporidium (oo)cysts in shellfish waters (Francavilla et al. 2012). It is probable that protozoa cannot survive longer than a few weeks in winter and a few days in summer as a result of the higher water temperatures coupled with solar irradiation and shallow lagoon depths (Francavilla et al. 2012). Conversely, summertime peak in incidence of AGIs depicts higher air temperatures influence the fate and replication of bacterial and protozoan pathogens (Galway et al. 2014) while temperatures were shown to be irrelevant in the survival of E. coli in a lagoon environment used for bathing (Schernewski et al. 2012).

Increases in mean or maximum air temperature have been shown to be significantly associated with verotoxigenic Escherichia coli (VTEC) outbreak occurrence at a national level (O'Dwyer et al. 2016) and with increasing AGI cases in a rainfall-dominated regime (Galway et al. 2015). Increased cases of bacterial AGI (Brubacher et al. 2020) and bacillary dysentery (Zuo et al. 2021) have also been associated with increased temperatures. Increased risks of gastroenteritis outbreaks (Herrador et al. 2021), campylobacteriosis (Trájer et al. 2022), coliform bacteria (Herrador et al. 2021) and E. coli (Sanchez et al. 2021) are associated with high temperatures. In addition, preceding air temperatures of 1–3 weeks had an influence on occurrence of giardiasis cases (Brunn et al. 2019). While giardiasis was positively associated with temperature, cryptosporidiosis was negatively associated (Britton et al. 2010).

From relative risk estimates of temperature and WBDs, O'Dwyer et al. (2016) discovered that for every degree Celsius temperature increase, the probability of a waterborne VTEC outbreak increases by a factor of 1.370, while the number of monthly AGI cases increased by 11.5% for a 1 °C increase in monthly mean temperature (95% CI: 4.5–18.6%) (Galway et al. 2015). For three bacterial AGIs (campylobacteriosis, VTEC and salmonellosis), the combined relative risk was 1.1 (95% CI: 1.02–1.21) for every 1 °C in annual mean temperature (Brubacher et al. 2020).

While not widespread in the articles reviewed, a couple of studies identified an association between solar radiation and WBD risk. It is imperative to highlight these associations given observed and projected changes for solar radiation intensity in some regions due to changes in climate. For instance, increasing solar radiation was associated with a decline in E. coli levels in the harvest of baby spinach from the irrigated waters (Allende et al. 2017), a decline in E. coli and FIB concentrations at Toronto (Sanchez et al. 2021) and German (Herrig et al. 2019) beaches, and reduced Cryptosporidium and Giardia (oo)cysts in harvested shellfish from shallow lagoons (Francavilla et al. 2012). Sunlight exposure had a significant influence on the decay of pathogens such as Cryptosporidium (Ahmed et al. 2023) and E. coli (Oladeinde et al. 2014) in cow pats. In particular, ultraviolet radiation (100–400 nm in wavelengths) is an effective disinfection treatment for drinking water, preventing pathogen replication and rendering them non-infective (Monis et al. 2014).

Evidence on links between winds and WBD risk are documented and require greater elucidation based on observed and projected changes in wind speed and direction for some regions due to climate change. For instance, wind speed has been shown to be associated with bacillary dysentery (Zuo et al. 2021) and prevailing wind conditions (wind speed and wind direction) have been identified as significant in the transport and concentrations of E. coli in larger surface water bodies (Schernewski et al. 2012; Eregno et al. 2018). Even though the possible mechanisms behind this are not established, there are strong interactive effects of flow conditions and wind speed on transport of pathogens and risk of WBDs, especially in the summer months. This suggests a potential influence of wind and flow direction (hydroclimatic drivers) on the transport of pathogens, such as E. coli along shorelines which favour pathogen accumulations. In addition, when an E. coli emission scenario was undertaken, the model simulations showed a high potential of E. coli concentrations in the bathing area of the beach above the threshold (500 CFU/100 ml) of sufficient bathing water quality as stipulated by the European Union directive (2006/7/EC) (Schernewski et al. 2012) for most wind speeds simulated. After emission into the lagoon, E. coli were not transported toward the open lagoon but rather along the shoreline creating an enabling environment for the bacteria's survival and subsequent accumulation and increased concentrations of the bacteria (Schernewski et al. 2012).

Increased frequencies and intensities of extreme events such as drought, landslip-related events and bushfires have been observed in some temperate regions in recent decades as a consequence of climate change. It is therefore important to elucidate observed associations between these events and WBD risk. For instance, drought and post-drought sampling periods showed a significant influence on the presence and concentrations of E. coli and total coliforms in private groundwater wells in the Republic of Ireland (O'Dwyer et al. 2021). Localized specific mechanisms (such as local specific tank density, local livestock density and localized preferential flow) appear to be the predominant source contamination of E. coli in sampled groundwater supplies, indicating a significant hydrodynamic shift from intrinsic and infrastructural mechanisms previously reported (Hynds et al. 2012; O'Dwyer et al. 2018), and reflecting the potential adaptation of the pathogen within vulnerable groundwater systems (O'Dwyer et al. 2021).

Conversely, landslip-related events and bushfires had significant influence on the concentrations of Cryptosporidium in a large drinking water reservoir in Australia (Bertone et al. 2016). Landslip events such as landslides can be a dominant influence on the turbidity of the reservoir following heavy rainfall events, usually occurring in winter with limiting dilution capacities. Ashes from bushfires around the catchment washed into the reservoir through runoff following heavy rain events increased levels of colour and turbidity in the reservoir. While colour and turbidity in the reservoir may increase Cryptosporidium levels, other potential factors such as presence of intensive livestock or a sewage treatment plant overflow can increase pathogen concentrations (Bertone et al. 2016).

While the focus of this review was on climate-related drivers of WBDs, about 30% of studies included non-climatic drivers such as health-seeking behaviour, individual behaviour or risk perception, local changes in land use, seasonal outdoor activities, tourism, densely populated areas, failed sewer or stormwater systems, growing season of agricultural crops and local soil types or properties as part of their assessments. As such, a section was included in this review to capture the non-climatic drivers that were presented within studies whose primary focus was climate-related. These non-climate-related factors, though not the focus of this study, were stated as secondary drivers that were interrelated with climate-related factors in some of the included studies. Due to the increasing climate-change impacts in some temperate regions coupled with the extent of human reactions and responses in its mitigation, it is pertinent to highlight these other drivers. Health-seeking behaviours of those who live in urban areas and are socioeconomically advantaged show some link to increased risk of WBDs (Forbes et al. 2021), though results may be biased due to lack of access to those who are disadvantaged. Another non-climate-related factor is individual behaviour or risk perception that affects risk of exposure and therefore contraction of disease. In assessing rural residents' perceived contamination risks on their private wells from the effects of surface water flooding prevalent in the Republic of Ireland, Musacchio et al. (2021) discovered that rural residents respond in different ways: some had a belief that floods were unlikely to recur and they were unperturbed about the event, especially non-agricultural residents; and many (over 75%) residents who had experienced flood were less concerned about taking proactive measures against flood. However, proactive measures were adopted (<50% of respondents) by those who had prior experiences of well contamination during floods. In addition, Andrade et al. (2019) observed that Health Beliefs Model (HBM) may be more cost-effective when dealing with participants that were unfamiliar with the event than Risks-Attitudes-Norms-Abilities-Self-regulation (RANAS) framework when predicting residents' perceived risks to WBD after such extreme events.

Local changes in land use are associated with WBD risk (Forbes et al. 2021). Weather-related changes in local landscapes and disruption to daily activities for rural Inuit communities (such as hunting, fishing, foraging, and travelling) affect physical, mental and emotional wellbeing and are associated with increased risk of WBDs (Cunsolo Willox et al. 2012). Late summer periods are also critical periods for AGI incidence as a result of increased involvement of populations in outdoor activities influenced by higher temperatures (Galway et al. 2014; Jalliffier-Verne et al. 2015) such as barbecues, swimming and canoeing (David et al. 2014), and playing on floodplains (Gertler et al. 2015). Tourism is also associated with increased risk of E. coli during the summer months as anthropogenic pollution is increased near the coast due to large numbers of tourists (Gentili et al. 2014; Bojarczuk et al. 2018) and when beach sand sediments are stirred during bathing (Schernewski et al. 2012).

Densely populated areas are associated with increased WBD risk through faecal contamination from pressurized municipal wastewater plants or leaky septic tanks (Gentili et al. 2014; Mohammed et al. 2019b; Forbes et al. 2021; O'Dwyer et al. 2021). Potential human faecal contamination of separated stormwater systems is also associated with WBD risk (Petterson et al. 2016). Finally, risk of E. coli is influenced by the growing season of agricultural crops where organic and livestock manure applied on heaps and fields can be potentially transferred to irrigation trenches or groundwater (Allende et al. 2017; Buckerfield et al. 2019), and local subsoil type or soil properties were associated with the presence of E. coli (Invik et al. 2019; O'Dwyer et al. 2021).

To date, relatively few studies have analyzed an ensemble of climate change models and climatic scenarios in predicting future rates of waterborne pathogens/diseases. Elucidating on projections for future risks of WBDs in temperate regions is pertinent due to observed changes and lack of quantification of potential impacts of climate change. For projection rates of cryptosporidiosis and giardiasis, and Cryptosporidium and Campylobacter, Chhetri et al. (2019) and Sterk et al. (2016), respectively examined changes to extreme precipitation. In both studies, precipitation amounts are expected to increase, coupled with an increase in rainfall over winter. As a result, overall risk of infection is predicted to increase. While Cryptosporidium infection was projected to be higher than that of Campylobacter (Sterk et al. 2016), annual rates of giardiasis and cryptosporidiosis are expected to increase by approximately 16% by the 2080s (Chhetri et al. 2019), coinciding with the projected period of extreme precipitation and wetter months (Coffey et al. 2016). Simulated scenarios also show a substantial impact of rainfall on E. coli contamination of baby spinach in irrigated waters through bacterial transfer onto the plants from contaminated manure-amended soil when splashing water on the fields (Allende et al. 2017).

Results of simulations that focused on coupling climate change with microbial fate and transport at the catchment scale indicate that increased intensity and frequency of winter rainfall are strongly associated with a period of increased risk of enteric pathogens (Coffey et al. 2016). On the other hand, for the winter and spring seasons, projected E. coli concentrations in two Norwegian streams (for 2045 and 2075) may remain similar to current levels (Mohammed et al. 2019a). While high concentrations are currently prevalent in the summer, lower concentrations of E. coli may be observed in both Norwegian streams in the future during the summer (Mohammed et al. 2019a). In addition, a 20-month simulation for analyzing long-term variations of E. coli in the lake showed similarities in concentrations in measured and simulated water temperatures at different lake depths (Hong et al. 2021). Elevated bacterial AGI rates in some ecological zones are projected to expand in range by the 2080s (Brubacher et al. 2020).

Other studies used streamflow changes estimated by climate change models to project future waterborne pathogens/diseases. In cases where flows are projected to increase, there is a subsequent increase in mean E. coli concentrations. This can be especially acute during spring floods (Jalliffier-Verne et al. 2015). Furthermore, future scenarios integrating high streamflow, decrease in combined sewers outlets (CSOs) and population growth are expected to result in increased mean E. coli concentrations by about 87% at drinking intakes downstream (Jalliffier-Verne et al. 2017).

The following climate emission scenarios were adopted for the projections: Relative Concentration Pathway (RCP): RCP 4.5 (Coffey et al. 2016; Trájer et al. 2022), RCP 8.5 (Coffey et al. 2016; Chhetri et al. 2019; Mohammed et al. 2019a; Trájer et al. 2022), RCP 2.6 and 6.0 (Trájer et al. 2022). Conversely, an ensemble of climate models used among the included studies including the following: Canadian Regional Climate Model (CRCM) coupled with the Canadian Global Climate Model (CGCM-3) following the Special Report on Emissions Scenarios (SRES)-A2 (Jalliffier-Verne et al. 2015); Community Climate Systems Model version 4 (CCSM4) (Trájer et al. 2022); a variety of regional climatic models, the EC-Earth international consortium and other global projections to downscale global data to the local scale in Ireland (Coffey et al. 2016); four different climate change scenarios developed by the Royal Netherlands Meteorological Institute (KNMI) for the Netherlands (Sterk et al. 2016); and Brubacher et al. (2020) who based his climate projections on previous work done by Wang et al. (2012) where 20 combinations of global climate models and SRES (A2, B1, A1B), were selected. It is noteworthy that observed differences in scenarios used in studies are highlighted in the literature.

Thirty-three models were applied in included studies to investigate the influence of climatic factors on waterborne pathogens/diseases and simulate climatic scenarios, including data-driven models, process-based hydrodynamic and water quality models and, in a few instances, a combination of both. The possibility of combining data-driven models and process-based models to assess WBD risk is getting increasing attention in recent studies (Mohammed et al. 2022) and brings together a consortium of vast transdisciplinary knowledge that can help solve complex societal issues (Hong et al. 2021). A comparison of singular or combined models used to account for pathogens or diseases is depicted in Supplementary material, Table S2. With the exclusion of source, other processes of fate, transport, exposure and risk of illness were investigated in two studies (see Supplementary material, Table S2). One of the studies combined a process-based model, Generalized Environmental Modelling System for Surface Waters Hydrodynamic model (GEMSS-HDM) and a data-driven model, Quantitative Microbial Risk Assessment (QMRA) to investigate the influence of climate on waterborne pathogens (Eregno et al. 2016). On the other hand, the other study coupled Seasonal Water Yield module (SWYM), a process-based model, and Geodetector model (GM), a data-driven model, in their investigation (Zuo et al. 2021). Most of the other studies focused on either assessing exposure and illness or the processes of source, fate and transport of pathogens and not both, which may explain the current gaps in knowledge (see Supplementary material, Table S2).

This study explores how climate change impacts relationships between climate-related factors and waterborne pathogens/diseases in temperate regions, examines how future changes in frequencies and durations of climate-related factors may affect future risks of WBDs in temperate regions and identifies models used to project future disease risks. From the included studies, complex relationships between climatic, hydrological and behavioural factors and WBD risk were analyzed. Identified drivers indicate a link between climate, hydrology and behaviour, and their resultant effects on WBD risk in a changing climate. In particular, heavy intensity precipitation, especially after prolonged dry periods, has been more associated with increased levels of E. coli (the most studied pathogen) and have been commonly evident in catchment outlets (Gentili et al. 2014; Buckerfield et al. 2019). Interestingly, bacteria accumulate and survive for extended periods of time on shores of rivers (Schernewski et al. 2012; Gentili et al. 2014), undisturbed temperate soils (Byappanahalli et al. 2003) and riparian sediments (Solo-Gabriele et al. 2000; Byappanahalli et al. 2003). Therefore, when heavy rainfall events occur, there is a tendency of rain–runoff events to transport bacteria from the river shores in the headwaters to the catchment outlets (Buckerfield et al. 2019), together with the influence of streamflow and wind speed/direction (Schernewski et al. 2012). Evidence of increased transport and concentrations of E. coli in surface water associated with prevailing wind conditions, along with flow conditions supports this (Schernewski et al. 2012).

Pathogen survival in river sediments and its potential for resuspension is an area for future research (Brunn et al. 2019). In essence, associations between soils along river banks and sediment transport, including their relationship to E. coli should be explored to establish processes involved in the association (Schernewski et al. 2012; Allende et al. 2017). With current and projected increases in precipitation intensity in some parts of temperate regions, the likelihood of E. coli levels increasing in source waters may continue. The literature has further identified that the link between rainfall events and protozoan contamination or quantification is limited (Aguirre et al. 2016) and more studies are recommended. In addition, the relationship between extreme precipitation after prolonged dry periods and protozoan pathogens should be further explored to establish the relationship (Chhetri et al. 2017), as well as associations between extreme weather events such as floods or droughts and waterborne infections more generally (Boudou et al. 2021; Herrador et al. 2021). Future studies are also needed to explore the possible roles of longer scale Earth system processes (such as El Niño Southern Oscillations (ENSO)) on bacterial and protozoan AGI, as already explored for cholera (Galway et al. 2014).

Even though streamflow was the most reported hydrological variable associated with WBD risk among included studies, more studies are still needed to better establish their relationship. This is pertinent especially with the early snowmelt being experienced in much of Canada and western USA because of a changing climate. Due to complex relationships that exist among climate, environment and pathogens, further studies should include more environmental factors so as to achieve a better understanding of the relationship and identify significant drivers of infection risk (Brubacher et al. 2020; Zuo et al. 2021).

Mean or maximum air temperatures are shown to be associated with increased WBD risk in temperate regions. With projected increases in temperatures in temperate regions, there is the likelihood that increasing trends in WBD risk may continue. However, this increment is reported more so for bacterial pathogens than for protozoan pathogens/illnesses. For protozoan illnesses (giardiasis and cryptosporidiosis), further studies are needed to establish their relationship with temperature (Chhetri et al. 2019). Further studies are also recommended in establishing associations between water temperatures and risk of WBDs.

Complex interrelationships between climatic, hydrological, topographic and human/animal activity factors need to be further explored, including social processes, to provide a better understanding of significant drivers of infection risk (Brubacher et al. 2020). Some studies that incorporated social processes indicate the influence of faecal contamination of surface waters from large agriculture-intensive catchments (Allende et al. 2017) as a driver of infection risk. Waterfowls were also identified as significant channels of bacterial contamination in the water (Petterson et al. 2016). Increased E. coli levels were also more reported during longer summer seasons in some regions due to changes in climate through increased tourism and recreation. High-densely populated areas and industrialized regions, which are centres of attraction, may also draw large tourists, especially to coastal communities experiencing changes in climate where the risk of faecal contamination of municipal wastewater is greater (Gentili et al. 2014). Populations are also drawn to increased outdoor activities such as swimming and canoeing during prolonged summers due to climate change in some regions whereby potential of ingestion rates is high (Corsi et al. 2016), and stirring of beach sand sediments is increased, influencing a rise in WBD risk. Assessing climate–pathogen relationships by season may allow for better identification and capture of short-term processes (Brubacher et al. 2020). In particular, for seasonal activities, it is best to examine the exposure frequency of such seasonal activities from appropriate seasons to obtain more effective options in the control and prevention of infections (David et al. 2014).

The future health impacts of climate change on WBD risks will vary over spatial and temporal scales, and may also depend on changing environmental and socioeconomic conditions (IPCC 2022). For projections of infection rates among included studies, climatic scenarios reveal that extreme precipitation is expected to increase to the 2080s with winters becoming wetter and infection rates are projected to increase to the 2080s, especially for E. coli, giardiasis and cryptosporidiosis/Cryptosporidium. Simulated future climate reveals that, as temperatures continue to rise to the 2080s, bacterial AGI rates will also increase in some biogeoclimatic zones (Brubacher et al. 2020). Other scenarios project seasonal variations in streamflows are associated with projected mean increases in E. coli concentrations (Jalliffier-Verne et al. 2017, 2015).

Observed projections from these studies should, however, be interpreted with caution because of the possibility of varied results at different spatial and temporal scales. The time scale at which data used for research on climate-disease relationships (such as aggregating monthly counts of infections) may either mask or unmask underlying patterns of associations which may, otherwise, be revealed if analysis of relationships were at daily or weekly scales (Galway et al. 2015; Corsi et al. 2016; Brunn et al. 2019). Spatial scales can also be a limitation, as what may be significant at a smaller spatial scale may not be at a larger scale (Zuo et al. 2021). Dynamics of infectious diseases, especially WBDs, can become clearer when, for instance, the influence of rainfall is examined at a large-scale level such as postal area level to identify short-term rainfall patterns (Forbes et al. 2021).

In identifying associations between WBD risk and climate-related factors coupled with making projections in a changing climate, studies used more data-driven models (such as QMRA and distributed lag non-linear regression model (DLNM)) to assess exposure and illness than process-based models (such as Soil and Water Assessment Tool (SWAT) and GEMSS-HDM) to analyze processes of source, fate and transport of pathogens (see Supplementary material, Table S2). However, two studies that integrated both categories of models to assess fate and transport of pathogens, and analyze exposure and illness (see Supplementary material, Table S2), had the advantage of simulating different processes over direct monitoring of pathogen concentrations. For instance, inputs for a QMRA through continuous monitoring of pathogen concentrations in six Norwegian beaches can be costly and unrealistic; therefore, adding hydrodynamic model simulation can be a potential substitution for monitoring (Eregno et al. 2016). Considering associated uncertainties, integrating hydrodynamic modelling with QMRA potentially contributed to an improved understanding of spatial–temporal spread of pathogens after rainfall events at recreational beaches and an improved QMRA technique for recreational waters (Eregno et al. 2016). Through the combination of a spatial statistical analysis, Geodetector model (data-driven model) and an ecosystem model, InVEST (process-based model) that integrated multi-source data (including meteorological, hydrological, topographic and socioeconomic variables) to investigate interactive relationship, one study was able to explore the influence of hydrological factors on bacillary dysentery disease in three different geographical regions in Anhui province, China (Zuo et al. 2021). Therefore, future studies should explore integrating data-driven models and process-based models so that simulation of different processes (hydrological, meteorological, topographical or social) can be integrated to better identify impacts and projections of future disease risk (Hong et al. 2021; Mohammed et al. 2022).

This review provides the current state of knowledge on the influence of climate-change impacts on associations between climate-related factors and WBD risk through a synthesis of evidence of 75 studies in temperate regions. While rainfall and heavy intensity rainfall were the most reported climatic drivers in many locations across temperate regions, streamflow was the most reported hydrological driver of WBD risk in Europe. Current and projected increases in frequency and intensity of precipitation and warming, especially in Europe, may be important. WBD risk is likely to increase in temperate regions with increases in mean monthly or annual temperatures and extreme precipitation events due to changes in climate, especially for E. coli (the most studied bacteria), Cryptosporidium (the most studied protozoa), cryptosporidiosis (the most reported WBD) and giardiasis (the second most reported WBD). However, these projected results should be interpreted with caution as there could be changes in WBD risk if other factors such as increases in source water quality monitoring and having more effective water treatment plants are reinforced.

Furthermore, climate-change impacts on summer and autumn seasons linked with social processes through increased recreation, tourism or agricultural activities are associated with increasing levels of WBD risk. Finally, combining data-driven models and process-based models may be a probable solution to accurately predict future disease risk with further model improvements. The ability to investigate spatial and temporal dynamics of pathogen risk in freshwater sources and its ability to be applied repeatedly under dynamic conditions, such as land use changes, improve our understanding in designing more effective intervention measures for tackling potential hazards of climate-change impacts on WBD risk in temperate regions. Findings from this review and identified areas for future research can serve as a guide to relevant decision makers in the design of strategies and policies that will inform decisions on better management, prevention and control of current and future WBD risk in temperate regions.

This work was supported by the Global Water Futures and Global Institute for Water Security.

E.A.S. conceptualized, wrote, reviewed, and edited the article. Z.G. worked on project administration and data curation. C.J.S and A.P. conceptualized, reviewed, and edited the article.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.