https://orcid.org/0000-0003-3192-8576
Figures
Abstract
The frequency and severity of heatwaves are increasing, posing challenges for understanding their effects on host-parasite dynamics. Especially, our understanding of the role of specific heatwave attributes in shaping disease outcomes remains limited. In this study, the Daphnia magna–Ordospora colligata host-parasite system, a widely used model for environmentally transmitted diseases, was used to investigate heatwave attributes. The amplitude and duration of heatwaves were manipulated across four baseline temperatures and four distinct time points relative to host exposure to the pathogen. This design resulted in 64 unique heatwave treatments, with O. colligata fitness (measured as prevalence and proliferation) recorded at the individual level in temperature-controlled water baths. Results show that heatwaves can alter parasite burden up to 13-fold, whereby amplitude, duration, and timing can interact with baseline temperature. Our results reveal complex interactions between heatwave attributes and baseline temperature, emphasising that heatwaves have context-dependent effects on parasite prevalence and proliferation. Additionally, when compared to other types of temperature variation (for example, cold snaps), heatwaves behave differently. While specific effects may vary across systems, these results demonstrate that interactions between heatwave attributes and baseline temperature can drive substantial variation in infection outcomes. These findings highlight the challenges and complexities involved in understanding and predicting how climate change and extreme weather events may influence disease dynamics in the context of global change. This underscores the need to incorporate thermal fluctuations into disease ecology models, as host-parasite responses to climate extremes are unlikely to be uniform across taxa.
Citation: McCartan N, O’Keeffe F, Zhang G, Luijckx P (2025) Impact of heatwave amplitude, duration, and timing on parasite fitness at different baseline temperatures. PLOS Clim 4(6): e0000632. https://doi.org/10.1371/journal.pclm.0000632
Editor: Noureddine Benkeblia, University of the West Indies, JAMAICA
Received: February 5, 2025; Accepted: April 28, 2025; Published: June 4, 2025
Copyright: © 2025 McCartan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets generated and code used to analyse during the current study are available in the GitHub repository: https://github.com/niamhmccartan/heatwave (copy archived at https://zenodo.org/records/15114093) McCartan N., O’Keeffe F., Zhang G., and Luijckx P. Impact of heatwave amplitude, duration, and timing on parasite fitness at different baseline temperatures. Zenodo Repository https://doi.org/10.5281/zenodo.15114093 (2025).
Funding: This work was supported by Science Foundation Ireland Frontiers for the Future (19/FFP/6839 to PL and NMcC). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. NMcC received a PhD stipend from this funding.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Climate change is increasingly recognised as a pivotal driver of alterations in human health [1], biodiversity [2], and ecosystem changes [3]. The rapid anthropogenic change in climate can disrupt host-parasite dynamics [4–6] as it has been well established that the host-parasite relationship can be modified by temperature [7–11]. Indeed, a recent study reported that 58% of human pathogenic diseases have been aggravated by climate change [12]. Temperature can influence host susceptibility by altering host homeostasis [13], immunity [14], and behaviour [15]. For example, in marine ecosystems, elevated sea temperatures can increase the prevalence of coral diseases by impairing coral immune functions and promoting pathogen growth [16]. Moreover, temperature changes can directly affect the parasite by modifying its infectivity [17], proliferation [18], and survival in the environment [19]. For example, the parasitic trematode Ribeiroia ondatrae infecting an amphibian host displayed an increase in host penetration but a decrease in infectivity and survival outside of the host depending on temperature [20]. Thus, the outcome of host-parasite relationships in response to temperature can be complex [18, 21, 22]. Yet, climate change is not only a linear increase in mean temperature [23], but it also modifies temperature variability, leading to altered temperature fluctuations and more frequent extreme weather [24].
Climate change is expected to increase the frequency and magnitude of anomalous weather events, such as heatwaves, cold snaps, and droughts [25]. Indeed, the frequency, intensity, and duration of heatwaves have accelerated globally since the 1950s [26,27]. Like rising mean temperatures, these temperature shifts caused by extreme and fluctuating weather can influence the survival of both the host and the parasite [18]. For example, thermal fluctuations at the lower thermal limits of malaria will increase infection rates, while decreasing infection rates at the upper thermal limits [28]. Moreover, climatic extremes are expected to result in biodiversity loss [29], changes in coevolutionary trajectories [21], and shifts in ecosystem structure [30] as the host-parasite relationship is altered. For instance, abnormally cooler temperatures in Oregon led to sea star wasting disease in the sea star Pisaster ochraceus, which resulted in the collapse of this keystone predator, leading to a trophic cascade [31]. Regardless of parasite exposure, thermal extremes are expected to influence life history traits [32], including development [33], ageing [34], and reproductive output [35]. Furthermore, temperature variability can disrupt transmission pathways, alter host susceptibility, and shift parasite life cycles in multi-host parasite systems [36,37]. Such effects are particularly relevant for vector-borne and environmentally transmitted pathogens, where hosts and parasites may respond differently to thermal fluctuations. However, not all thermal fluctuations or extreme weather events are the same; different attributes of such temperature events can lead to alterations in physiological responses.
The amplitude and duration of temperature fluctuations and extreme weather events are known to have variable outcomes on host-parasite interactions [38–43]. For example, immune responses in ectothermic hosts can vary depending on the frequency and severity of thermal fluctuations, with some species exhibiting immune suppression under prolonged heat stress while others show resilience through plastic physiological adjustments [44,45]. Extreme temperature events such as heatwaves can also modulate immune function, either by suppressing immune defences or altering energetic trade-offs [46,47]. Such immune shifts may have cascading effects on disease susceptibility, influencing parasite establishment and proliferation in hosts experiencing thermal stress. Additionally, the temporal scale of multiple stressors (e.g., heatwave attributes) should be considered to add realistic complexity when exploring disease outcomes [48]. Indeed, the timing of heat stress relative to infection can influence which aspects of the host-parasite interaction are most affected as parasites may trigger immune responses or other physiological changes [49]. Moreover, predictive models incorporating thermal performance curves suggest that climate-driven shifts in disease prevalence will depend on complex interactions between temperature, parasite fitness, and host susceptibility [5, 50]. These models emphasise the importance of considering non-linear thermal effects when forecasting future disease risks under climate change. To better understand how these complex interactions will alter the host and parasite, it is important to consider how specific temperature events, like heatwaves, may directly affect physiological responses of the host. Still, while some papers [41, 51, 52] have focused on the effect of heatwaves on the host-parasite relationship, few have comprehensively examined the effect of varying heatwave attributes across multiple temperatures [38, 41].
Here, we investigated the impact of varying heatwave attributes (that is timing, amplitude, and duration) across a broad spectrum of baseline temperatures (representing the mean temperature maintained throughout the experimental period). We used the Daphnia magna – Ordospora colligata host-pathogen model, a popular model system for environmentally transmitted diseases (characterised by horizontal transmission mechanisms and infection dynamics governed by mass action principles) [53]. The studies by Kunze et al. (2022) and McCartan et al. (2024) were the basis used for the experimental design. Although both studies found alternative patterns in parasite burden in the same Daphnia-Ordospora system depending on the heatwave/cold snap treatment (i.e., a pulse of extreme temperature), they did not specifically manipulate heatwave attributes. Building upon this observation, we hypothesised that not all heatwave treatments may result in the same outcome in parasite fitness, and this may depend on the baseline temperature, amplitude and duration of the heatwave, and timing in relation to disease exposure. In fact, parasite performance can be reduced as a result of thermal stress when temperatures become too high (as suggested by the Thermal Stress Hypothesis [54]), while shifts in temperatures may lead to increased parasite performance according to the Thermal Variability Hypothesis [9]. This hypothesis suggests that smaller organisms, such as parasites, have a faster metabolic rate than their larger hosts and consequently will acclimate more quickly to temperature changes compared to their hosts (as outlined by the Metabolic Theory of Ecology [55]). Moreover, as thermal performance curves are inherently non-linear, small changes in temperature can result in disproportionately large changes in host or parasite performance due to Jensen’s Inequality [38]. In addition, as shifts in temperature are known to distinctly affect various host and parasite life history traits [11, 20, 41, 54, 56], this may further complicate how parasite life cycles, disease dynamics, and epidemics are impacted by global change. Overall, this means that the effects of temperature fluctuations on the host-parasite relationship may be complicated, as performance can be affected by the variability in temperature, altering expectations regarding disease outcome [57]. As anticipated, our findings revealed that heatwaves can induce unexpected and intricate shifts in parasite infection prevalence and proliferation, which will impede the prediction of host-pathogen relationships under climate change.
Materials and Methods
Two experiments were conducted to explore the effect of heatwaves (i.e., pulse heat treatments) on parasite fitness. Experiment 1 focused on the effect of baseline (constant) temperature, timing, amplitude, and duration on infection prevalence and burden in Daphnia magna infected with Ordospora colligata. Experiment 2 focused on the repeatability of results using one baseline temperature and one amplitude but altered the timing and duration of heatwave treatments.
Study system
The freshwater crustacean Daphnia magna (genotype Fi-OER-3–3) was infected with its microsporidian parasite Ordospora colligata (isolate 3) to simulate 64 heatwave treatments (both host and parasite were originally sampled from Tvärminne, Finland, and maintained in the lab for over 10 years). Daphnia magna is a widely studied freshwater crustacean found across the northern hemisphere, which clonally reproduces via parthenogenesis, allowing for genotypic consistency [58]. They are a popular model system for the study of host-pathogen interactions [53] due to their fast generation time, their well-known ecology, and their key role in aquatic ecosystems [59]. Ordospora colligata is a microsporidian microparasite (i.e., an infectious disease) which is transmitted horizontally to the host via filter feeding. It invades the upper gut epithelium, where it forms intracellular clusters of 32–64 spores, which burst and are shed into the water via faeces [58]. We may expect to see differences in the performance of the antagonists in this system in response to heatwaves, given that Daphnia have a wider thermal range of 6-33.3°C compared to Ordospora’s narrower range of 11.8-29.7°C [7].
Temperature-controlled water baths
Water baths were used to ensure the water temperature was accurately controlled throughout the experiment. Each bath was fitted with an aquarium chiller (Hailea HC-150A, DC300, or DC750), an aquarium heater (EHEIM JÄGER 300W) and pumps (Micro-Jet Oxy and Oase Optimax 500) to ensure even temperature distribution, and a programmable controller (Inkbird ITC-308 or ITC-310T) for regulation. Temperatures were monitored using a HOBO logger (HOBO UA-001–08) and checked daily. Heatwaves were simulated by moving individual microcosms directly to their assigned temperature bath, while the baths were maintained at a constant baseline temperature for the duration of the experiment. Acclimatisation to the heatwave treatment started immediately upon entering the bath, and microcosms took less than an hour to reach the target temperature. When the heatwave treatments were concluded, the microcosms were returned to their original position.
Experiment 1 setup
The methods used for this experiment were similar to those outlined by McCartan et al. (2024). A fully factorial experimental design was conducted where the timing, amplitude, and duration of heatwaves were altered across four baseline (constant) temperatures (14, 17, 20, and 23°C) using eighteen temperature-controlled water baths ranging from 14-29°C. Heatwave duration was varied by 3 days or 6 days (two durations), and heatwave amplitude was increased by +3°C or +6°C (two amplitudes). These baseline temperatures were chosen to remain within the thermal limits of Ordospora (11.8–29.7°C) while ensuring infection success and thermal stress at some baseline temperatures. Heatwave treatments increased temperatures by up to 6°C to reflect realistic warming scenarios [60], representing an extreme but plausible event that pushes conditions toward Ordospora’s upper thermal threshold. After the heatwave treatment was completed, treatments returned to their baseline temperature. The experiment commenced 10 days prior to exposure to Ordospora (day -10). Individual juvenile female Daphnia were added to allocated microcosms held in the water baths prior to the start of the first heatwave (day -10), and subsequent heatwave treatments occurred on day 0 (day of exposure to O. colligata), day 10, and day 20 (four timings) (Fig 1). In total, there were 64 heatwave treatments (4 baseline temperatures • 4 heatwave timings • 2 amplitudes • 2 durations), each with 15 replicates. A constant temperature treatment (amplitude 0°C) was included for each baseline temperature with additional constant controls at 26°C and 29°C. Likewise, uninfected (placebo) controls were included at each temperature to confirm that no unintended spread of Ordospora occurred.
TIMELINE; illustrates the timeline for the experiment starting at day -10 and terminating at day 27, exposure occurred on day 0, and heatwaves began on day -10, 0, 10, and 20. PREPARATION; preparation started three weeks prior to exposure when 10-12 mothers were added to 400 mL microcosms, 72 hours before the experiment began, offspring were collected and females were kept. INFECTION; a dose of ~60,000 Ordospora colligata spores was given to each exposed individual, while unexposed controls were given a placebo dose made of crushed-up uninfected individuals. BATH SETUP; each temperature had three baths, and each bath held four trays with 27 microcosms. Each bath contained 80 heatwave-treated individuals (five replicates per heatwave treatment), eight constant temperature treatments, and four uninfected controls. To simulate a heatwave, microcosms were moved between baths and then returned to the baseline bath when the heatwave finished. TREATMENTS; 64 treatments were included in the experiment (four baseline temperatures • four heatwave timings • two amplitudes • two durations). Additionally, at the heatwave-specific temperatures (26°C and 29°C), a tray held 27 microcosms but only two constant treatments and one uninfected control, the remaining microcosms contained no Daphnia, so these could be moved when the heatwave occurred. MAINTENANCE; maintenance and measurements were carried out between days -8 and day 27. Figure created on Biorender.com.
All treatments were organised into trays where a tray contained twenty-seven 100 mL microcosms (filled with 80 mL of Artificial Daphnia Medium (ADaM [61]). A single tray contained two constant temperature treatments, one uninfected control and five replicates of a single heatwave treatment. Each temperature had three baths, and each bath held four trays. One tray per bath corresponded to a single heatwave timing (i.e., all heatwave treatments in the same tray began a simulated heatwave treatment at the same time) (Fig 1). Trays were rotated every day in a clockwise direction within each bath. In total, there were 1176 animals, each placed in a separate microcosm (i.e., 960 heatwave treatments, 144 constant treatments and 72 uninfected controls).
Daphnia preparation and exposure to Ordospora
A group of 10–12 females were added to 400 mL microcosms with ~300 mL of ADaM, transferred twice a week, and fed ad libitum with batch culture algae (Scenedesmus sp.) to stimulate the production of offspring [62]. The juvenile offspring used to initiate the experiment were collected in a 72-hour period and sexed (using a dissecting microscope at 8x to 12x magnification). Only females were kept to remove variation in sex [58], which corresponds to the dominant state of the natural populations, which are female-biased. The experiment began on day -10 when Daphnia juveniles were individually added to a 100 mL microcosm filled with 50 mL of ADaM and a pinch of Cetyl alcohol to break the surface tension. The individuals remained unexposed until day 0 when Daphnia in the exposed treatments received a 1 mL dose of ADaM containing ~60,000 Ordospora spores. The spore dose was created by crushing ~3580 individuals with a known average spore burden, and the slurry was diluted to 1300 mL using ADaM. Uninfected control treatment Daphnia received a placebo dose made up by crushing uninfected Daphnia. Individuals were transferred to fresh ADaM every four days (to avoid the accumulation of offspring produced and waste products) and fed every two days (from 5 million algae/mL on day -10–12 million algae/mL by day 5, which remained constant until the end of the experiment). Day 0 also coincided with the start of a heatwave, which commenced after all Daphnia received either an infective or placebo dose.
Experiment 2 setup
A second experiment was carried out to test the reproducibility of the results. Here, timing and duration were altered at the same amplitude (+6°C) and a single baseline temperature (17°C). Heatwaves occurred 10 days prior to exposure, the day of exposure (day 0) and 10 days post-exposure. A heatwave occurring 20 days post-exposure was excluded to focus on timings before, during, and after exposure at equal intervals. Duration was also manipulated, and individuals were subjected to either a 3-day or a 6-day heatwave at 23°C (two durations). Constant temperature controls and uninfected individuals to control for accidental infections were also present in this experiment. In total, there were six treatments (3 heatwave timings • 2 durations • 1 amplitude). There were two baths at the baseline temperature (17°C) and two heatwave baths (23°C). Each baseline bath held four trays (with 27 microcosms in each tray), and each bath held 90 heatwave-treated individuals (thirty replicates per heatwave treatment), fourteen constant treatments, and three uninfected controls. In total, experiment 2 contained 248 individual Daphnia, each in an individual microcosm. The bath and microcosm setup was the same as in experiment 1. Likewise, infection preparation was similar, but ~1148 Daphnia were used to make an infective dose of ~60,000 spores in a 400 mL dilution, while uninfected individuals received a placebo dose. The feeding regime was also identical to experiment 1.
Measurement of parasite fitness for both experiments
To measure parasite fitness, the infection prevalence (i.e., presence or absence of spores) and burden (i.e., the number of spore clusters in the host) were examined in each individual. Daphnia were checked for infection upon natural death or upon termination at the end of the experiment (within five days of day 27). If infected, spore clusters (each cluster holding up to 64 individual spores) were counted using bright field or phase-contrast microscopy (with 400x magnification). For experiment 1, deaths occurring before day 9 (n = 98) were omitted due to the lack of accuracy in the diagnosis of infection status at this early stage. Any inconclusive infections (n = 35) and misidentified males (n = 2) were also omitted. For experiment 2, deaths occurring before day 7 (n = 25) were omitted due to the lack of accuracy in diagnosis, and inconclusive infections (n = 7) and males (n = 1) were also omitted.
Data analysis for both experiments
Analysis was performed using R version 4.0.3 [63]. Infection prevalence (i.e., presence or absence of infection), exposure (i.e., number of Ordospora spore clusters, including zeros for uninfected exposed individuals), and burden (i.e., the number of Ordospora spore clusters excluding zeros) were the response variables. Experiment 1 used baseline temperature, timing, amplitude, and duration as the explanatory variables, while Experiment 2 used experiment, timing, and duration as the explanatory variables. Infection prevalence and exposed data included all exposed individuals, while burden data included confirmed infections only. Exposure data can be found in the supplement (Tables D and E in S1 Text) and corresponded largely to the output of the model for ‘burden’. Furthermore, ‘bath’ was excluded as a variable as ‘temperature’ was treated as a continuous centred variable to reduce collinearity issues in experiment 1, and was not significant in experiment 2 (GLM, analysis of deviance χ2(2,275) p = 0.8618).
For experiment 1, baseline temperature was treated as a continuous centred variable to address collinearity concerns, with a cubic polynomial added to the burden model to accommodate the non-linear temperature response, while infection prevalence was modelled with a linear relationship. For variable selection and regression modelling, elastic net regression was applied using the “glmnet” package [64] as a Generalised Linear Model (GLM) with a Poisson distribution, a GLM with a negative binomial distribution (glm.nb), and a Generalised Linear Mixed Model (GLMM), all had problems with multicollinearity. Elastic net regression combines L1 (lasso) and L2 (ridge) regularisation by shrinking coefficients and preventing overfitting, thus mitigating the impact of collinearity. While elastic net regression is effective for handling multicollinearity and high-dimensional data, some limitations should be noted. It does not support random effects, so batch-specific variability (e.g., ‘bath’) could not be accounted for. Additionally, while it helps manage overfitting, elastic net may still have limitations in fully capturing complex random effects that might be present in the data. Therefore, while mixed-effects models, like GLMMs, were considered, ‘glmnet’ was chosen for its effectiveness in managing multicollinearity and high-dimensional data.
The ‘cv.glmnet’ function was used with a binomial distribution for infection prevalence. The regression fit (alpha = 0.5) was selected based on the lowest cross-validation error, while the regularisation strength (lambda = 0.0467) was chosen via cross-validation with one standard error ‘lambda.1se’. For exposed data, the ‘glmnet’ function was chosen with a regression fit (alpha = 0.85) based on the lowest AIC score. The regularisation strength of the model was also chosen by the lowest AIC score (lambda = 0.0050). The negative binomial family (theta = 0.7260) was selected to account for overdispersion. Burden also used the same ‘glmnet’ function and has a regression fit (alpha = 0.7) and regularisation strength (lambda = 0.0248) based on the lowest AIC score. A negative binomial family (theta = 1.3112) was also used in this model. Bootstrapping (50,000 iterations) was used for infection prevalence, exposure, and burden to assess the model stability and significance, resulting in the estimated coefficients (effect) and their 95% confidence intervals. Finally, feature importance was assessed for each factor by calculating the absolute value of the effect coefficients from the elastic net regression model. This approach allowed us to identify the most influential features in predicting the outcome (response variable). The process was performed separately for each polynomial degree, and the results were then aggregated to determine the overall importance of each variable. Custom contrasts were created with the “emmeans” package [65] to compare the average infection prevalence and burden of the heatwave treatments. For this, timing, amplitude, and duration were combined into a new variable ‘treatment’, and modelled against temperature for clarity of interpretation. The contrast p-values were then adjusted for multiple comparisons using the ‘Benjamini-Hochberg’ method [66].
Data analysis to compare experiments 1 and 2 focused on burden using a GLM with a negative binomial distribution (‘glm.nb’) to observe the individual effects of timing, duration, and experiment (the difference between experiments 1 and 2), with no interaction between the variables as these were not significant. An analysis of deviance (χ 2) was then carried out to assess the overall significance of the predictor variables. Finally, to observe the effect of specific treatments in experiments 1 and 2, the “emmeans” package was used to compare burden against the predictor variable ‘treatment’ (combined variables: experiment, timing, and duration). In other words, experiment 1 treatments were compared to one another, and experiment 2 treatments were compared, but all in the same model (Table K in S1 Text). Like with experiment 1, contrast p-values were adjusted for multiple comparisons using the ‘Benjamini-Hochberg’ method.
Results
Infection prevalence
All factors (i.e., the timing, amplitude, and duration of the heatwave, and baseline temperature) influenced infection prevalence. Timing of the heatwave (before, during, or after infection) led to variable outcomes (Fig 2). Infection prevalence was reduced in heatwaves beginning on the day of exposure (day 0) that had a strong amplitude of +6°C above the baseline temperature (most important predicting infection prevalence, linear effect, glmnet, effect = -0.44, 95% CI: -0.49 to -0.31 and feature importance = 0.44, Table 1). These heatwaves only had a single infection at 23°C (Fig 2B) when the heatwave co-occurred during parasite exposure. Indeed, here, the strong and short heatwave treatment (a heatwave of +6°C for 3 days) was different from the constant treatment (emmean, z(inf) = -3.848, p < 0.008, see Table C in S1 Text for p-values of all custom contrasts). Yet, when the same heatwave occurred 10 days prior to exposure, no reduction in infection prevalence was observed (Fig 2A). Furthermore, only a small reduction (not significant) was observed when a heatwave occurred 10- and 20-days post-exposure for the short 3-day heatwave with a strong amplitude of +6°C (Fig 2C and 2D). However, regardless of the timing of the heatwave, amplitude and duration also influenced the host-parasite interaction. The strong amplitude and long duration heatwave (heatwave of +6°C for 6 days) differentially influenced prevalence compared to other treatments depending on the baseline temperature and was important in predicting prevalence whereby prevalence decreased with increasing temperature (linear effect, glmnet, effect = -0.17, 95% CI: -0.25 to -0.06, Table 1). For example, at 23°C, a strong and long heatwave (heatwave of +6°C for 6 days) occurring 10 days post-exposure was different from the constant temperature treatment (emmean, z(inf) = -3.904, p < 0.008, see Table C in S1 Text, Fig 2C). Finally, the baseline temperature was also important for predicting infection prevalence (linear effect, glmnet, effect = -0.08, 95% CI: -0.14 to -0.02 and feature importance = 0.08, Table 1). Infection prevalence declined as baseline temperatures increased across all heatwave timings (Fig 2). The high feature importance of two complex 3-way interactions present in the infection prevalence model (i.e., indicating how much these variables contribute to the model’s predictions) indicates that baseline temperature, timing of the heatwave, amplitude, and duration all influence infection prevalence in the Daphnia-Ordospora system.
Each spore cluster contains 32-62 individual Ordospora. Error bars represent the standard error, and each dashed line is the linear fit of the elastic net regression (see Table A in S1 Text for statistical results). Grey; constant treatment, yellow; heatwave lasting 3 days where the temperature was increased by 3°C, orange; heatwave lasting 6 days where the temperature was increased by 3°C, pink; heatwave lasting 3 days where the temperature was increased by 6°C, purple; heatwave lasting 6 days where the temperature was increased by 6°C. A; heatwave treatments which began on day -10, B; heatwave treatments which began on day 0, C; heatwave treatments which began on day 10, and D; heatwave treatments which began on day 20. All factors (heatwave timing, amplitude, duration and baseline temperature) were important for predicting infection prevalence. A strong amplitude treatment (which increased temperature by 6°C) occurring on the day of exposure influenced infection prevalence depending on baseline temperature. Furthermore, the heatwave treatment, which increased temperature by 6°C for 6 days specifically, had distinct effects on infection prevalence depending on baseline temperature. Finally, baseline temperature, regardless of any heatwave attributes, was also important for predicting infection prevalence.
Burden
For parasite proliferation, the effects of a heatwave ranged from a 13.5-fold decrease in burden to a 2.4-fold increase (Fig 3, Fig 4). Specifically, individuals that experienced a heatwave with a baseline temperature of 23ºC, 10 days post-exposure, with an amplitude of +6ºC for 6 days had a lower burden (mean of 34 spore clusters) than the baseline constant (mean of 458 spore clusters) (Fig 4L), with only one successful infection. Conversely, a heatwave on the day of exposure (day 0) with an amplitude of +6ºC for 6 days and a baseline temperature of 14 ºC resulted in a 2.4-fold increase in burden compared to the baseline constant (mean of 472 vs 197 spore clusters, respectively) (Fig 4E). Thus, similar to infection prevalence, the number of spore clusters within the Daphnia depended on the timing of the heatwave, the amplitude and duration, and the baseline temperature (Fig 3). This is supported by the high feature importance of the complex 3-way and 2-way interactions (Table 2), which highlight the intricacy and context-specific nature of the outcomes in burden.
Each spore cluster contains 32-62 individual Ordospora. Error bars represent the standard error, and each dashed line is the cubic fit of the elastic net regression (see Table F in S1 Text for statistical results). The absence of error bars at any point indicates the presence of a single data point for the treatment. Grey; constant treatment, yellow; heatwave lasting 3 days where the temperature was increased by 3°C, orange; heatwave lasting 6 days where the temperature was increased by 3°C, pink; heatwave lasting 3 days where the temperature was increased by 6°C, purple; heatwave lasting 6 days where the temperature was increased by 6°C. A; heatwave treatments which began on day -10, B; heatwave treatments which began on day 0, C; heatwave treatments which began on day 10, and D; heatwave treatments which began on day 20. No infections occurred in the + 6ºC for 6 days, but the data point (0) was added for clarity in visualisation. Baseline temperature was important for predicting burden. The strong amplitude treatment (which increased temperature by 6°C) occurring 20 days post-exposure influenced burden, but the influence of this effect depended on baseline temperature. Regardless of heatwave timing, a short-duration heatwave (3 days) could also influence burden depending on baseline temperature. Specifically, a heatwave treatment, which increased temperature by 3°C for 3 days, had distinct effects on burden depending on baseline temperature.
Points are grouped by duration. The solid grey line represents the mean for constant treatments. Error bars represent the standard error; the light grey area represents the standard error for the constant treatments. An asterisk next to a solid black bracket (*[) indicates significant contrasts between the means of the heatwave treatments when compared to the constant treatment or between two single treatments. For all custom contrasts using emmeans, see Table H in S1 Text.
Baseline temperature influenced burden and resulted in the second-highest feature importance (quadratic effect, glmnet, effect = -19.40, 95% CI: -22.36 to -16.81 and feature importance of the combined polynomial = 19.40, Table 2). Overall, similar to infection prevalence, burden followed a classic temperature curve with the highest spore counts around the optimal temperature (~17°C) and lower parasite performance at the extremes (Fig 3). However, in heatwave treatments occurring after exposure (day 10 and 20), proliferation generally decreased compared to the constant treatment (Fig 3C and 3D). Here, with rising baseline temperature, the negative effect on burden accelerated, resulting in a non-linear decelerating relationship. Indeed, alternate baseline temperatures can result in differential outcomes between heatwave treatments. For example, for heatwaves occurring 20 days post-exposure, comparing the strong and long (+6°C for 6 days) and the weak and long (+3°C for 6 days) heatwaves at 14°C and 20°C led to differential outcomes (Fig 3D). At 14°C, the strong and long heatwave had a greater burden (although not significantly greater) than the weak and long heatwave (emmean, z(inf) = 0.904, p = 0.808, see Table H in S1 Text, Fig 4M). However, at 20°C, the opposite was true, whereby the weak and long heatwave had the greater burden (emmean, z(inf) = -3.01, p = 0.048, see Table H in S1 Text, Fig 4O). Moreover, while baseline temperature can influence the proliferation of Ordospora, the timing of the heatwave can lead to alternative outcomes.
Timing of the heatwave relative to parasite exposure influenced burden and affected parasite performance, leading to shifts in thermal performance compared to the constant treatments. For example, there was no difference in burden with any heatwaves occurring 10 days pre-exposure when compared to the baseline constant (Fig 3A, Table F in S1 Text). Yet, heatwaves occurring 20 days post-exposure resulted in a shift in optimal parasite performance towards the colder temperatures (effect = -0.54, 95% CI: -0.76 to -0.20 and feature importance of the combined polynomial = 0.54, Table 2, Fig 3D). Timing also interacted with baseline temperature and amplitude to influence proliferation. Specifically, at 20 days post-exposure, a strong amplitude (+6°C) heatwave resulted in a lower burden compared to the constant treatments at and above 20°C (linear effect, glmnet, effect = -14.65, 95% CI: -21.91 to -0.75 and feature importance of the combined polynomial = 25.52, Table 2, Fig 3D). For example, at 23°C, both the strong and short (+6°C for 3 days) and the strong and long (+6°C for 6 days) heatwaves had a lower burden than the constant (emmean, z(inf) = -5.933, p < 0.001 and emmean, z(inf) = -4.099, p = 0.001, respectably, see Table H in S1 Text, Fig 4P). Thus, the timing of the heatwave can influence parasite burden and result in complex interactions depending on the amplitude and duration of the heatwave and the baseline temperature at which it occurs.
Regardless of timing, amplitude and duration also influenced parasite proliferation. Indeed, burden was influenced by a short-duration (3 days) heatwave mediated by baseline temperature (cubic effect, glmnet, effect = 9.44, 95% CI: 3.66 to 14.78 and feature importance of the combined polynomial = 9.82, Table 2). For example, burden increased in a short-duration (3 days) heatwave occurring 10 days post-exposure at 17ºC, while a strong and long heatwave (+6ºC for 6 days) reduced burden (Fig 4J). In other words, a short-duration heatwave mediated by baseline temperature resulted in a non-linear response in burden where an increase in temperature accelerated burden at median temperature ranges and decelerated at others. More specifically, the weak and short heatwave (+3°C for 3 days) can influence burden depending on the baseline temperature (linear effect, glmnet, effect = 6.77, 95% CI: 0.13 to 12.38 and feature importance of the combined polynomial = 9.82, Table 2, Fig 3). For example, at 14°C, there was a 1.6-fold decrease in burden in the heatwave treatment (+3°C for 3 days) (mean of 126 spore clusters) compared to the constant treatment (mean of 197 clusters). Furthermore, certain combinations of timing, amplitude, and duration of heatwave treatments also resulted in differential outcomes in burden. A heatwave occurring at 14ºC on the day of exposure (day 0) with a weak amplitude and long duration (+3°C for 6 days) resulted in a 2.7 times higher spore burden than a weak and short heatwave (+3°C for 3 days) (emmean, z(inf) = 2.970, p = 0.048, see Table H in S1 Text, Fig 4E). Therefore, heatwave treatments can lead to differing outcomes in parasite fitness, and the impact can depend on the amplitude and duration of the heatwave as well as the timing and baseline temperature at which it occurs.
Experiment comparisons
Analysis of both experiments showed that both the timing and duration of the heatwave altered the Ordospora burden compared to constant temperature treatments. However, there was a difference between the burden in experiments 1 and 2 whereby the overall burden of experiment 1 was higher than experiment 2 by a mean of 422 spore clusters (GLM, analysis of deviance X2(1,276) p < 0.001) although there were no interactions (Fig 5, Table 3). Duration also influenced burden (GLM, analysis of deviance X2(1,272) p = 0.029), and although no individual treatments were different (no significant custom contrasts), a long heatwave (6 days) tended to have a lower burden compared to a short heatwave (3 days) (Fig 5). Finally, timing also influenced burden (GLM, analysis of deviance X2(1,273) p = 0.013), whereby no timing resulted in the same burden (Fig 5). Thus, both timing and duration were consistently important in influencing the outcome of spore burden.
Red boxplots indicate a 3-day heatwave, and blue boxplots indicate a 6-day heatwave. The grey dashed line shows the mean burden for the constant baseline treatment at 17°C, and light grey solid lines represent the upper and lower standard errors for the constant. All heatwaves had an amplitude of +6°C. Boxplots were separated by timing of heatwave (10 days pre-exposure, the day of exposure (day 0), and 10 days post-exposure). Within each timing, plots were separated by experiments: A; experiment 1 burden, and B; experiment 2 burden.
Discussion
The effects of heatwaves on host-parasite dynamics are shaped by multiple complex interactions involving baseline temperature, heatwave timing, amplitude, and duration. Our findings demonstrate that these factors can interact in intricate and often unpredictable ways, leading to context-dependent outcomes for both infection prevalence and parasite proliferation. This highlights the inherent complexity in understanding and predicting the impacts of heatwaves on disease dynamics and a need for deeper exploration into heatwave attributes and potential underlying mechanisms.
Effect of baseline temperature
Heatwaves can influence the performance of Ordospora colligata, but the pattern and direction of the effects are dependent on the baseline temperature at which the heatwave occurs. A strong and long heatwave (i.e., an increase in the amplitude of +6ºC above the baseline temperature for 6 days) occurring on the day of exposure (day 0), 10-, and 20-days post-exposure at high baseline temperatures (23ºC) can strongly reduce infection prevalence. Therefore, baseline temperature can greatly influence parasite fitness. Indeed, lower parasite performance at thermal extremes may be caused by thermal stress as the parasite is unable to survive outside of its thermal tolerance, as suggested by the Thermal Stress Hypothesis [54]. Since Ordospora has a narrower thermal tolerance (11.8-29.7°C) compared to its host (6-33.3°C) [7], the heatwaves approached the parasite’s upper thermal tolerance while the Daphnia were less affected. However, parasite thermal performance limits may differ, with some species exhibiting adaptive plasticity to alterations in temperature [67,68]. For example, malaria and trematodes can adjust their development to temperature variation [69]. Similarly, in marine trematode-amphipod interactions, temperature fluctuations have been shown to alter parasite transmission success, likely due to changes in host susceptibility and parasite survival [70]. If Ordospora lacks similar plasticity, it may be disproportionately affected by temperature extremes, leading to reductions in infection success during heatwaves. Thermal stress may also explain the reduction in burden for heatwaves after exposure (day 10 and day 20) with a strong amplitude treatment (+6ºC) (compared to the constant baseline temperature). A decrease in parasite performance near the thermal extremes was previously demonstrated by McCartan et al. (2024) and in other systems (e.g., butterfly-protozoa, mosquito-malaria, and frog-chytrid fungus systems [71–73]). For instance, in monarch butterflies (Danaus plexippus) infected with their protozoan parasite Ophryocystis elektroscirrha, infections were reduced at extreme temperatures due to the inability of the parasite to invade the host and replicate [71]. Yet, thermal stress may not account for all variations in response to baseline temperature.
The Thermal Mismatch hypothesis and Jensen’s Inequality could explain some of the variability in parasite fitness when Ordospora is exposed to heatwaves. The Thermal Mismatch Hypothesis [56] posits that differences in the optimal temperature ranges of the host and parasite, driven by temperature fluctuations, can shape their interactions, potentially giving an advantage to either the host or the parasite, resulting in context-specific outcomes that are complex [11]. For example, at the thermal extremes where the Daphnia’s thermal performance exceeds that of Ordospora, the parasite is at a disadvantage. Other studies that have measured host and parasite thermal performance found evidence for the Thermal Mismatch Hypothesis [67, 74]. For instance, the parasitoid Cotesia congregate will suffer more under increasing temperatures than its lepidopteran larval host, Manduca sexta, due to a difference in thermal tolerances [75]. Over evolutionary timescales, such thermal mismatches may alter host-parasite co-evolutionary trajectories [49]. Hosts under recurrent heat stress may evolve enhanced thermal resilience, potentially reducing parasite transmission over time [76]. Conversely, parasites that can rapidly adapt to warming conditions may gain a selective advantage, particularly in multi-host systems where thermal stress disrupts transmission pathways [77]. However, given that it is not possible to measure the thermal performance of Ordospora independently of its Daphnia host (as it cannot be grown in culture media), formally testing this hypothesis in the Daphnia-Ordospora system is currently not possible. Moreover, Jensen’s Inequality, which states that averaging over a nonlinear curve can lead to deviations and thus disproportionately large alterations to the host-parasite relationship, may also explain some of the observed differences [78]. The presence of multiple quadratic and cubic terms when Ordospora burden is modelled across temperature indicates that the parasite’s response to temperature is non-linear and heatwaves can thus lead to larger than expected increases or decreases in parasite performance. Overall, that heatwaves can result in a shift in thermal performance depending on the baseline temperature at which the heatwave occurred has been shown in other systems [40, 73, 79], highlighting the importance of including fluctuations in disease models. For example, the fitness of the moth Spodoptera littoralis is overestimated at higher temperatures when thermal fluctuations are ignored [80]. Thus, heatwaves can affect the parasite’s success depending on the baseline temperature; however, timing, duration, and amplitude also influence parasite fitness when temperature is varied.
Effect of Timing
No heatwave timings resulted in the same outcome for either infection prevalence or burden. For example, while no infections occurred in a strong amplitude (+6ºC) at 23ºC on the day of exposure (day 0), the same treatments had a 100% infection rate when the heatwaves occurred 10 days prior to exposure. Moreover, the timing of the heatwave played a role in both experiments, with similar impacts, but different overall burdens. Other systems have highlighted the importance of heatwave timing, such as the parasitoid (Eretmocerus hayati), where performance decreased depending on the heatwave timing and duration [81]. Additionally, a long duration (6 days) heatwave occurring on the day of exposure (day 0) led to an increased burden at lower temperatures (14ºC and 17ºC) as Ordospora may have had enough time in the optimal temperature to successfully establish as higher temperatures lead to increased feeding rates and greater contact rate between Daphnia and Ordospora [17]. Other systems have shown that thermal extreme events can alter host heat tolerance [49], parasite survival [82] and affect host susceptibility [83,84]. For example, short-term heatwaves can disrupt parasite development within Anopheles mosquitoes, where transient high temperatures delay or block oocyst maturation, ultimately reducing transmission potential [69]. However, while we see a reduction in burden in Ordospora when exposure occurred during a heatwave (day 0), this was only the case in heatwaves with a high baseline temperature (23ºC) and larger amplitude (+6ºC). Here, increased mortality associated with thermal stress [54] or thermal mismatch [56] between host and parasite may make it harder for the parasite to establish. In other systems, like Crithidia bombi in bumblebees, no differences in infection were found when a heatwave occurred after exposure, potentially because any negative effects resulting from infection were counteracted and masked by the host’s susceptibility [83]. Yet, not all differential effects of the timing of a heatwave relative to parasite exposure may be explained by heat stress and another aspect to consider is how temperature may interact with host immunity [46,47].
Heat stress can temporally shape immune function in invertebrates, with effects varying depending on the severity and duration of exposure [85]. In fish-parasite interactions, increased temperatures can either enhance or suppress immune responses, depending on the severity and duration of the thermal stress [86]. Indeed, elevated temperatures can suppress immune defences, leading to increased susceptibility, for example, heatwaves induced long-lasting immune disorders in three-spined sticklebacks (Gasterosteus aculeatus) [14]. Moreover, heat-induced immune priming can also enhance resistance [87,88]. For example, in the mealworm beetle (Tenebrbrio beetle) brief exposure to elevated temperatures enhances resistance to bacterial infections through upregulated immune defences [89]. Here, a heatwave occurring 20 days post-exposure could have increased Daphnia immune function, which would aid in the clearance of Ordospora, leading to decreased spore burdens, especially with a strong amplitude (+6 ºC) heatwave. This is seen in other systems, for example, the Pacific white shrimp (Litopenaeus vannamei) where elevated temperatures post-infection activated Heat Shock Factor 1, leading to the upregulation of antimicrobial peptides that enhanced antiviral defences [90]. These temperature-driven immune modulations across taxa may explain why the timing of heatwaves relative to infection is critical in determining infection success, as seen in other host-parasite systems. Thus, the timing of a heatwave relative to infection can result in context-dependent outcomes, with varying effects for the same heatwave treatment depending on its timing relative to exposure.
Effect of Amplitude and Duration
Alternate amplitudes and durations of a heatwave can result in differential outcomes for parasite prevalence and proliferation, indicating that the impact of heatwaves on parasite fitness is context-dependent. Moreover, that the duration of a heatwave is key, is highlighted by the replicability of its importance in both experiments. While strong heatwaves (+6ºC) can reduce both infection and burden with increasing baseline temperatures (depending on the heatwave timing), a short-duration heatwave (3 days) can increase burden (at specific baseline temperatures). For example, when a heatwave occurred 10 days post-exposure at 17ºC, the burden increased from a mean of 950–1260 spore clusters. A 4-fold increase in the burden of Ordospora following a heatwave at 16ºC was also observed by Kunze et al (2022). Short-term fluctuations have also been known to increase infection in the Daphnia-Ordospora system. For example, endemic infection prevalence increased in the Daphnia-Ordospora system with short fluctuations of ±6ºC [38]. Furthermore, an increase in disease following thermal fluctuations has been observed in several study systems (e.g., human-parasite [69], animal-parasite [22], and plant-parasite [91]). For instance, short-term temperature fluctuations are believed to increase dengue virus infections in Aedes aegypti compared to large temperature fluctuations [42]. Here, the increase in burden following temperature fluctuations could be explained by the Thermal Variability Hypothesis [10]. This hypothesis posits that parasites should acclimatise faster to changing environments due to their mass-specific differences in metabolic rates, giving them an advantage over their larger host (as stated by the Metabolic Theory of Ecology [55]). Overall, both amplitude and duration are important in many systems for predicting the fitness of a parasite [92–94]. For instance, the amplitude of the heatwave was important in the host-feeding parasitoid Eretmocerus hayati [95], while heatwave duration affected the life history and immunity of Lymnaea stagnalis snails when exposed to trematodes, as prolonged duration heatwaves decreased host performance [96]. Conversely, heatwaves have also been known to reduce infection. For example, the chytrid fungus (Batrachochytrium dendrobatidis) reduced infection in frogs when subjected to pulse heat treatments [73]. Thus, both amplitude and duration are important to consider when predicting the fitness of Ordospora, particularly under the influence of climate change.
Conclusions
In conclusion, all factors (baseline temperature, timing of the heatwave, amplitude, and duration) influenced parasite fitness, leading to variable outcomes in the Daphnia-Ordospora system. Moreover, heatwaves behave differently compared to other forms of temperature variation. For example, cold snaps increase burden with increasing temperatures in the Daphnia-Ordospora system, whereas heatwaves decrease burden [43]. Overall, at high baseline temperatures, different combinations of amplitude and duration can decrease fitness, while at lower temperatures, only smaller fluctuations can increase burden, highlighting the complexity with added temperature variation. Our findings highlight the complex and context-dependent nature of heatwave impacts on parasite fitness.
Considering the complexity of the impact of heatwaves on Ordospora, and given Daphnia’s role as a key organism in their ecosystem [53], our results may have widespread impacts on ecology. For example, another parasite of Daphnia magna has been suggested to induce trophic cascades by reducing host densities, leading to increased algae densities and decreased water clarity [97]. Moreover, our study highlights the importance of including temperature variation in future models and its impact on disease dynamics. For example, it is believed that 69.3% of COVID-19 cases in the summer of 2022 could have been avoided if there had been no heatwaves [98]. Therefore, future research should be conducted to explore the dynamics between temperature variation and parasite fitness as climate change continues to result in increased frequency and intensity of heatwaves and other anomalous weather events [27]. A greater insight into how temperature variation affects host-parasite dynamics will aid in predicting ecological impacts, as parasites can affect host population dynamics [99], biodiversity [97], and ecosystem functioning [100]. Since heatwaves can either suppress or enhance parasite fitness depending on their characteristics [101, 102], identifying and incorporating the mechanisms underlying these effects in epidemiological models is essential for accurately assessing risks to wildlife and human health. Conservation efforts should also account for these effects, as parasite burdens may shift unpredictably under changing climates [103], further underscoring the need for refined predictive models and targeted mitigation strategies to assess the broader ecological consequences of climate-driven disease dynamics. Future studies should also consider the potential for adaptive plasticity and co-evolutionary dynamics across a wider range of host-parasite systems to better understand the broader implications of heatwaves on disease spread.
Supporting information
S1 Text. Table A: coefficient estimates and feature importance of infection prevalence in Daphnia magna infected with Ordospora colligata.
Calculated using elastic net regression, bootstrapping and feature importance. Coefficient Effect (Effect) represents the estimated coefficient of the ‘glmnet’ elastic net regression model. Positive values indicate a positive relationship between the variable and infection prevalence, while negative values indicate a negative relationship (i.e., a positive value means the explanatory variables increased prevalence). Effect sizes represent the strength of these relationships, with larger absolute values indicating stronger effects. Lower (Lower CI) and Upper (Upper CI) Confidence Intervals indicate the range (95th percentile) from bootstrapping; non-significant values have CIs encompassing 0, suggesting no important relationship to the response variable. Temperature was modelled as a linear (L) relationship. Feature importance (Imp) was calculated by taking the absolute value of the effect. Table B: estimated marginal means (emmeans) for infection prevalence in Daphnia magna infected with Ordospora colligata. The estimates were derived using a Generalized Linear Model (GLM) with baseline temperature as a factor and a treatment variable (timing, amplitude, and duration combined), and a binomial family. The presentation on the log scale is used to provide a more appropriate representation of the data. The ‘Inf’ degrees of freedom indicate the model’s high flexibility, which arises from the complex nature of the model itself. Table C: custom contrasts for infection prevalence ‘emmeans’ comparing treatments at different baseline temperatures in Daphnia magna Infected with Ordospora colligata. The ‘emmeans’ for each treatment were compared to the constant equivalent at the same temperature or to the alternate amplitude or duration at the same heatwave timing. ‘Inf’ degrees of freedom indicate the model’s complexity, making quantification challenging for estimated marginal means. Significant contrasts are bolded. Table D: coefficient estimates, and feature importance of exposed Daphnia magna infected with Ordospora colligata. Exposed included all individuals who received an infective dose. Calculated using elastic net regression, bootstrapping and feature importance. Coefficient Effect (Effect) represents the estimated coefficient of the ‘glmnet’ elastic net regression model. Positive values indicate a positive relationship between the variable and infection prevalence, while negative values indicate a negative relationship (i.e., a positive value means the explanatory variables increased prevalence). Effect sizes represent the strength of these relationships, with larger absolute values indicating stronger effects. Lower (Lower CI) and Upper (Upper CI) Confidence Intervals indicate the range (95th percentile) from bootstrapping; non-significant values have CIs encompassing 0, suggesting no important relationship to the response variable. Temperature was a cubic polynomial, therefore, to account for non-linearity and is therefore separated by each degree (linear [L], and quadratic [Q]). Feature importance (Imp) was calculated by taking the absolute value of the effect. Table E: combined feature importance for exposed Daphnia magna infected with Ordospora colligata. The feature importance was calculated by getting the absolute value of the coefficient for each variable from the coefficient estimates and then the three degrees per polynomial (linear, quadratic, and cubic) were combined. Table F: coefficient estimates, and feature importance of burden in Daphnia magna infected with Ordospora colligata. Exposed included all individuals who received an infective dose. Calculated using elastic net regression, bootstrapping and feature importance. Coefficient Effect (Effect) represents the estimated coefficient of the ‘glmnet’ elastic net regression model. Positive values indicate a positive relationship between the variable and infection prevalence, while negative values indicate a negative relationship (i.e., a positive value means the explanatory variables increased prevalence). Effect sizes represent the strength of these relationships, with larger absolute values indicating stronger effects. Lower (Lower CI) and Upper (Upper CI) Confidence Intervals indicate the range (95th percentile) from bootstrapping; non-significant values have CIs encompassing 0, suggesting no important relationship to the response variable. Temperature was a cubic polynomial, therefore, to account for non-linearity and is therefore separated by each degree (linear [L], and quadratic [Q]). Feature importance (Imp) was calculated by taking the absolute value of the effect. Table G: estimated marginal means (emmeans) for burden (number of spore clusters) in Daphnia magna infected with Ordospora colligata. The estimates were derived using a Generalized Linear Model (GLM) with baseline temperature as a factor and a treatment variable (timing, amplitude, and duration combined), and a negative binomial family. The presentation on the log scale is used to provide a more appropriate representation of the data. The ‘Inf’ degrees of freedom indicate the model’s high flexibility, which arises from the complex nature of the model itself. NA indicates lack of infected individuals in a treatment. Table H: custom contrasts for burden (number of spore clusters) ‘emmeans’ comparing treatments at different baseline temperatures in Daphnia magna Infected with Ordospora colligata. The ‘emmeans’ for each treatment were compared to the constant equivalent at the same temperature or to the alternate amplitude or duration at the same heatwave timing. ‘Inf’ degrees of freedom indicate the model’s complexity, making quantification challenging for estimated marginal means. Significant contrasts are bolded. NA indicates a lack of infected individuals in a treatment. Table I: combined feature importance for burden with importance in Daphnia magna infected with Ordospora colligata. The feature importance was calculated by getting the absolute value of the coefficient for each variable from the coefficient estimates and then the three degrees per polynomial (linear, quadratic, and cubic) were combined. Table J: estimated marginal means (emmeans) for burden (number of spore clusters), in Daphnia magna infected with Ordospora colligata combining two experiments. The estimates were derived using a Generalized Linear Model (GLM) with baseline temperature as a factor and a treatment variable (timing, amplitude, and duration combined), and a negative binomial family. The presentation on the log scale is used to provide a more appropriate representation of the data. The ‘Inf’ degrees of freedom indicate the model’s high flexibility, which arises from the complex nature of the model itself. NA indicates a lack of infected individuals in a treatment. Exp: experiment 1 (A) or experiment 2 (B). Table K: custom contrasts for burden (number of spore clusters) ‘emmeans’ comparing treatments at different baseline temperatures in Daphnia magna Infected with Ordospora colligata. The ‘emmeans’ for each treatment were compared to the constant equivalent at the same temperature or to the alternate amplitude or duration at the same heatwave timing. ‘Inf’ degrees of freedom indicate the model’s complexity, making quantification challenging for estimated marginal means. Exp: experiment 1 (A) or experiment 2 (B).
https://doi.org/10.1371/journal.pclm.0000632.s001
(DOCX)
Acknowledgments
The authors thank Dieter Ebert and Jürgen Hottinger for the provision of the biological materials, Alison Boyce for technical assistance in creating the water baths and maintenance of electrical equipment, and Peter Mc Cartan, Sean Hoare and Whitney Parker for helping with empirical work.
References
- 1. McMichael AJ, Woodruff RE, Hales S. Climate change and human health: present and future risks. Lancet. 2006;367(9513):859–69. pmid:16530580
- 2. Nunez S, Arets E, Alkemade R, Verwer C, Leemans R. Assessing the impacts of climate change on biodiversity: is below 2 °C enough? Clim Change. 2019;154(3):351–65.
- 3. Malhi Y, Franklin J, Seddon N, Solan M, Turner MG, Field CB, et al. Climate change and ecosystems: threats, opportunities and solutions. Philos Trans R Soc Lond B Biol Sci. 2020;375(1794):20190104. pmid:31983329
- 4. Brunner FS, Eizaguirre C. Can environmental change affect host/parasite-mediated speciation?. Zoology (Jena). 2016;119(4):384–94. pmid:27210289
- 5. Paull SH, Johnson PTJ. Experimental warming drives a seasonal shift in the timing of host-parasite dynamics with consequences for disease risk. Ecol Lett. 2014;17(4):445–53. pmid:24401007
- 6. Hance T, van Baaren J, Vernon P, Boivin G. Impact of extreme temperatures on parasitoids in a climate change perspective. Annu Rev Entomol. 2007;52:107–26. pmid:16846383
- 7. Kirk D, Jones N, Peacock S, Phillips J, Molnár PK, Krkošek M, et al. Empirical evidence that metabolic theory describes the temperature dependency of within-host parasite dynamics. PLoS Biol. 2018;16(2):e2004608. pmid:29415043
- 8. Kirk D, Luijckx P, Jones N, Krichel L, Pencer C, Molnár P, et al. Experimental evidence of warming-induced disease emergence and its prediction by a trait-based mechanistic model. Proc Biol Sci. 2020;287(1936):20201526. pmid:33049167
- 9. Rohr JR, Raffel TR. Linking global climate and temperature variability to widespread amphibian declines putatively caused by disease. Proc Natl Acad Sci U S A. 2010;107(18):8269–74. pmid:20404180
- 10. Raffel TR, Halstead NT, McMahon TA, Davis AK, Rohr JR. Temperature variability and moisture synergistically interact to exacerbate an epizootic disease. Proc Biol Sci. 2015;282(1801):20142039. pmid:25567647
- 11. Rohr JR, Cohen JM. Understanding how temperature shifts could impact infectious disease. PLoS Biol. 2020;18(11):e3000938. pmid:33232316
- 12. Mora C, McKenzie T, Gaw IM, Dean JM, von Hammerstein H, Knudson TA, et al. Over half of known human pathogenic diseases can be aggravated by climate change. Nat Clim Chang. 2022;12(9):869–75. pmid:35968032
- 13. Seppälä O, Jokela J. Immune defence under extreme ambient temperature. Biol Lett. 2011;7(1):119–22. pmid:20610417
- 14. Dittmar J, Janssen H, Kuske A, Kurtz J, Scharsack JP. Heat and immunity: an experimental heat wave alters immune functions in three-spined sticklebacks (Gasterosteus aculeatus). J Anim Ecol. 2014;83(4):744–57. pmid:24188456
- 15. Labaude S, Rigaud T, Cézilly F. Host manipulation in the face of environmental changes: Ecological consequences. Int J Parasitol Parasites Wildl. 2015;4(3):442–51. pmid:26835252
- 16. Harvell D, Jordán-Dahlgren E, Merkel S, Rosenberg E, Raymundo L, Smith G. Coral disease, environmental drivers, and the balance between coral and microbial associates. Oceanogr. 2007;20:172–95. https://doi.org/10.5670/oceanog.2007.91
- 17. Shocket MS, Vergara D, Sickbert AJ, Walsman JM, Strauss AT, Hite JL, et al. Parasite rearing and infection temperatures jointly influence disease transmission and shape seasonality of epidemics. Ecology. 2018;99(9):1975–87. pmid:29920661
- 18. Claar DC, Wood CL. Pulse Heat Stress and Parasitism in a Warming World. Trends Ecol Evol. 2020;35(8):704–15. pmid:32439076
- 19. Dias SRC, Dos Santos Lima W. Effect of temperature on activity of third-stage larvae of Angiostrongylus vasorum. Parasitol Res. 2012;110(4):1327–30. pmid:21861062
- 20. Paull SH, LaFonte BE, Johnson PTJ. Temperature‐driven shifts in a host‐parasite interaction drive nonlinear changes in disease risk. Global Change Biology. 2012;18(12):3558–67.
- 21. Wolinska J, King KC. Environment can alter selection in host-parasite interactions. Trends Parasitol. 2009;25(5):236–44. pmid:19356982
- 22. Raffel TR, Romansic JM, Halstead NT, McMahon TA, Venesky MD, Rohr JR. Disease and thermal acclimation in a more variable and unpredictable climate. Nature Clim Change. 2012;3(2):146–51.
- 23. Denny M. The fallacy of the average: on the ubiquity, utility and continuing novelty of Jensen’s inequality. J Exp Biol. 2017;220(Pt 2):139–46. pmid:28100801
- 24. van der Wiel K, Bintanja R. Contribution of climatic changes in mean and variability to monthly temperature and precipitation extremes. Commun Earth Environ. 2021;2(1).
- 25.
Leahy P, Gonzalez L, Hickey K, Kiely G, Allen M, Nowbakht P. Climatt: tools for climate change attribution of extreme weather events. 2021.
- 26. Perkins-Kirkpatrick SE, Lewis SC. Increasing trends in regional heatwaves. Nat Commun. 2020;11(1):3357. pmid:32620857
- 27.
Seneviratne SI, Zhang X, Adnan M, Badi W, Dereczynski C, Luca AD, et al. Weather and climate extreme events in a changing climate. In: Masson-Delmotte VP, Zhai A, Pirani SL, Connors C, Péan S, Berger N, et al., editors. Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press; 2021. p. 1513–766. https://doi.org/10.1017/9781009157896.013
- 28. Paaijmans KP, Heinig RL, Seliga RA, Blanford JI, Blanford S, Murdock CC, et al. Temperature variation makes ectotherms more sensitive to climate change. Glob Chang Biol. 2013;19(8):2373–80. pmid:23630036
- 29. Kohl WT, McClure TI, Miner BG. Decreased temperature facilitates short-term sea star wasting disease survival in the keystone intertidal sea star Pisaster ochraceus. PLoS One. 2016;11(4):e0153670. pmid:27128673
- 30. Byers JE. Effects of climate change on parasites and disease in estuarine and nearshore environments. PLoS Biol. 2020;18(11):e3000743. pmid:33232311
- 31. Menge BA, Cerny-Chipman EB, Johnson A, Sullivan J, Gravem S, Chan F. Sea star wasting disease in the keystone predator Pisaster ochraceus in Oregon: insights into differential population impacts, recovery, predation rate, and temperature effects from long-term research. PLoS One. 2016;11(5):e0153994. pmid:27144391
- 32. Ma C-S, Ma G, Pincebourde S. Survive a warming climate: insect responses to extreme high temperatures. Annu Rev Entomol. 2021;66:163–84. pmid:32870704
- 33. Stahlschmidt ZR. Warm and thermally variable incubation conditions reduce embryonic performance and carry over to influence hatchling tradeoffs. J Therm Biol. 2024;124:103946. pmid:39265502
- 34. Burraco P, Orizaola G, Monaghan P, Metcalfe NB. Climate change and ageing in ectotherms. Glob Chang Biol. 2020;26(10):5371–81. pmid:32835446
- 35. Sales K, Vasudeva R, Dickinson ME, Godwin JL, Lumley AJ, Michalczyk Ł, et al. Experimental heatwaves compromise sperm function and cause transgenerational damage in a model insect. Nat Commun. 2018;9(1):4771. pmid:30425248
- 36. Barber I, Berkhout BW, Ismail Z. Thermal change and the dynamics of multi-host parasite life cycles in aquatic ecosystems. Integr Comp Biol. 2016;56(4):561–72. pmid:27252219
- 37. Cohen JM, Civitello DJ, Venesky MD, McMahon TA, Rohr JR. An interaction between climate change and infectious disease drove widespread amphibian declines. Glob Chang Biol. 2019;25(3):927–37. pmid:30484936
- 38. Krichel L, Kirk D, Pencer C, Hönig M, Wadhawan K, Krkošek M. Short-term temperature fluctuations increase disease in a Daphnia-parasite infectious disease system. PLoS Biol. 2023;21(9):e3002260. pmid:37683040
- 39. Dallas T, Drake JM. Fluctuating temperatures alter environmental pathogen transmission in a Daphnia-pathogen system. Ecol Evol. 2016;6(21):7931–8. pmid:30128141
- 40. Duncan AB, Gonzalez A, Kaltz O. Stochastic environmental fluctuations drive epidemiology in experimental host-parasite metapopulations. Proc Biol Sci. 2013;280(1769):20131747. pmid:23966645
- 41. Kunze C, Luijckx P, Jackson AL, Donohue I. Alternate patterns of temperature variation bring about very different disease outcomes at different mean temperatures. Elife. 2022;11:e72861. pmid:35164901
- 42. Lambrechts L, Paaijmans KP, Fansiri T, Carrington LB, Kramer LD, Thomas MB, et al. Impact of daily temperature fluctuations on dengue virus transmission by Aedes aegypti. Proc Natl Acad Sci U S A. 2011;108(18):7460–5. pmid:21502510
- 43. McCartan N, Piggott J, DiCarlo S, Luijckx P. Cold snaps lead to a 5-fold increase or a 3-fold decrease in disease proliferation depending on the baseline temperature. BMC Biol. 2024;22(1):250. pmid:39472912
- 44. Rollins-Smith LA, Le Sage EH. Heat stress and amphibian immunity in a time of climate change. Philos Trans R Soc Lond B Biol Sci. 2023;378(1882):20220132. pmid:37305907
- 45. Ferguson LV, Kortet R, Sinclair BJ. Eco-immunology in the cold: the role of immunity in shaping the overwintering survival of ectotherms. J Exp Biol. 2018;221(Pt 13):jeb163873. pmid:29967267
- 46. Leicht K, Jokela J, Seppälä O. An experimental heat wave changes immune defense and life history traits in a freshwater snail. Ecol Evol. 2013;3(15):4861–71. pmid:24455121
- 47. Catalán TP, Wozniak A, Niemeyer HM, Kalergis AM, Bozinovic F. Interplay between thermal and immune ecology: effect of environmental temperature on insect immune response and energetic costs after an immune challenge. J Insect Physiol. 2012;58(3):310–7. pmid:22019347
- 48. Jackson MC, Pawar S, Woodward G. The temporal dynamics of multiple stressor effects: from individuals to ecosystems. Trends Ecol Evol. 2021;36(5):402–10. pmid:33583600
- 49. Hector TE, Gehman A-LM, King KC. Infection burdens and virulence under heat stress: ecological and evolutionary considerations. Philos Trans R Soc Lond B Biol Sci. 2023;378(1873):20220018. pmid:36744570
- 50. Molnár PK, Sckrabulis JP, Altman KA, Raffel TR. Thermal performance curves and the metabolic theory of ecology-a practical guide to models and experiments for parasitologists. J Parasitol. 2017;103(5):423–39. pmid:28604284
- 51. Milazzo A, Giles LC, Zhang Y, Koehler AP, Hiller JE, Bi P. Heatwaves differentially affect risk of Salmonella serotypes. J Infect. 2016;73(3):231–40. pmid:27317378
- 52. Schreven SJJ, Frago E, Stens A, de Jong PW, van Loon JJA. Contrasting effects of heat pulses on different trophic levels, an experiment with a herbivore-parasitoid model system. PLoS One. 2017;12(4):e0176704. pmid:28453570
- 53. Ebert D. Daphnia as a versatile model system in ecology and evolution. Evodevo. 2022;13(1):16. pmid:35941607
- 54. Paull SH, Raffel TR, LaFonte BE, Johnson PTJ. How temperature shifts affect parasite production: testing the roles of thermal stress and acclimation. Functional Ecology. 2015;29(7):941–50.
- 55. Brown JH, Gillooly JF, Allen AP, Savage VM, West GB. Toward a metabolic theory of ecology. Ecology. 2004;85(7):1771–89.
- 56. Cohen JM, Venesky MD, Sauer EL, Civitello DJ, McMahon TA, Roznik EA, et al. The thermal mismatch hypothesis explains host susceptibility to an emerging infectious disease. Ecol Lett. 2017;20(2):184–93. pmid:28111904
- 57. Bernhardt JR, Sunday JM, Thompson PL, O’Connor MI. Nonlinear averaging of thermal experience predicts population growth rates in a thermally variable environment. Proc Biol Sci. 2018;285(1886):20181076. pmid:30209223
- 58.
Ebert D. Ecology, epidemiology, and evolution of parasitism in Daphnia. Bethesda (MD): National Library of Medicine. 2005.
- 59. Sarnelle O. Daphnia as keystone predators: effects on phytoplankton diversity and grazing resistance. Journal of Plankton Research. 2005;27(12):1229–38.
- 60. Witze A. Extreme heatwaves: surprising lessons from the record warmth. Nature. 2022;608(7923):464–5. pmid:35927493
- 61. Klüttgen B, Dülmer U, Engels M, Ratte HT. ADaM, an artificial freshwater for the culture of zooplankton. Water Research. 1994;28(3):743–6.
- 62. Kilham SS, Kreeger DA, Lynn SG, Goulden CE, Herrera L. COMBO: a defined freshwater culture medium for algae and zooplankton. Hydrobiologia. 1998;377(1/3):147–59.
- 63.
RStudio Team. RStudio: Integrated development environment for R. Boston, MA: RStudio, PBC; 2022.
- 64. Friedman J, Hastie T, Tibshirani R. Regularization paths for generalized linear models via coordinate descent. J Stat Soft. 2010;33(1).
- 65.
Russell V. Lenth. emmeans: Estimated Marginal Means, aka Least-Squares Means 2022. Available from: https://CRAN.R-project.org/package=emmeans
- 66. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 1995;57(1):289–300.
- 67. Gehman A-LM, Hall RJ, Byers JE. Host and parasite thermal ecology jointly determine the effect of climate warming on epidemic dynamics. Proc Natl Acad Sci U S A. 2018;115(4):744–9. pmid:29311324
- 68. Aleuy OA, Kutz S. Adaptations, life-history traits and ecological mechanisms of parasites to survive extremes and environmental unpredictability in the face of climate change. Int J Parasitol Parasites Wildl. 2020;12:308–17. pmid:33101908
- 69. Paaijmans KP, Blanford S, Bell AS, Blanford JI, Read AF, Thomas MB. Influence of climate on malaria transmission depends on daily temperature variation. Proc Natl Acad Sci U S A. 2010;107(34):15135–9. pmid:20696913
- 70. Studer A, Poulin R. Differential effects of temperature variability on the transmission of a marine parasite. Mar Biol. 2013;160(10):2763–73.
- 71. Ragonese IG, Sarkar MR, Hall RJ, Altizer S. Extreme heat reduces host and parasite performance in a butterfly-parasite interaction. Proc Biol Sci. 2024;291(2015):20232305. pmid:38228180
- 72. Blanford JI, Blanford S, Crane RG, Mann ME, Paaijmans KP, Schreiber KV, et al. Implications of temperature variation for malaria parasite development across Africa. Sci Rep. 2013;3:1300. pmid:23419595
- 73. Greenspan SE, Bower DS, Webb RJ, Roznik EA, Stevenson LA, Berger L, et al. Realistic heat pulses protect frogs from disease under simulated rainforest frog thermal regimes. Functional Ecology. 2017;31(12):2274–86.
- 74. Nowakowski AJ, Whitfield SM, Eskew EA, Thompson ME, Rose JP, Caraballo BL, et al. Infection risk decreases with increasing mismatch in host and pathogen environmental tolerances. Ecol Lett. 2016;19(9):1051–61. pmid:27339786
- 75. Moore ME, Hill CA, Kingsolver JG. Developmental timing of extreme temperature events (heat waves) disrupts host-parasitoid interactions. Ecol Evol. 2022;12(3):e8618. pmid:35342573
- 76. Blanford S, Thomas MB, Pugh C, Pell JK. Temperature checks the Red Queen? Resistance and virulence in a fluctuating environment. Ecology Letters. 2002;6(1):2–5.
- 77. Cable J, Barber I, Boag B, Ellison AR, Morgan ER, Murray K, et al. Global change, parasite transmission and disease control: lessons from ecology. Philos Trans R Soc Lond B Biol Sci. 2017;372(1719):20160088. pmid:28289256
- 78. Ruel J, Ayres M. Jensen’s inequality predicts effects of environmental variation. Trends Ecol Evol. 1999;14(9):361–6. pmid:10441312
- 79. Beck-Johnson LM, Nelson WA, Paaijmans KP, Read AF, Thomas MB, Bjørnstad ON. The importance of temperature fluctuations in understanding mosquito population dynamics and malaria risk. R Soc Open Sci. 2017;4(3):160969. pmid:28405386
- 80. Bagni T, Siaussat D, Maria A, Fuentes A, Couzi P, Massot M. Fitness under high temperatures is overestimated when daily thermal fluctuation is ignored. J Therm Biol. 2024;119:103806. pmid:38335848
- 81. Zhang Y-B, Zhang G-F, Liu W-X, Wan F-H. Continuous heat waves change the life history of a host-feeding parasitoid. Biological Control. 2019;135:57–65.
- 82. Marcus E, Dagan T, Asli W, Ben-Ami F. Out of the “host” box: extreme off-host conditions alter the infectivity and virulence of a parasitic bacterium. Philos Trans R Soc Lond B Biol Sci. 2023;378(1873):20220015. pmid:36744562
- 83. Tobin KB, Mandes R, Martinez A, Sadd BM. A simulated natural heatwave perturbs bumblebee immunity and resistance to infection. J Anim Ecol. 2024;93(2):171–82. pmid:38180280
- 84. Porras MF, Navas CA, Agudelo-Cantero GA, Santiago-Martínez MG, Loeschcke V, Sørensen JG, et al. Extreme heat alters the performance of hosts and pathogen. Front Ecol Evol. 2023;11.
- 85. Murdock CC, Paaijmans KP, Cox-Foster D, Read AF, Thomas MB. Rethinking vector immunology: the role of environmental temperature in shaping resistance. Nat Rev Microbiol. 2012;10(12):869–76. pmid:23147703
- 86. Macnab V, Barber I. Some (worms) like it hot: fish parasites grow faster in warmer water, and alter host thermal preferences. Global Change Biology. 2011;18(5):1540–8.
- 87. Kurtz J, Franz K. Innate defence: evidence for memory in invertebrate immunity. Nature. 2003;425(6953):37–8. pmid:12955131
- 88. Browne N, Surlis C, Kavanagh K. Thermal and physical stresses induce a short-term immune priming effect in Galleria mellonella larvae. J Insect Physiol. 2014;63:21–6. pmid:24561359
- 89. Herren P, Hesketh H, Dunn A, Meyling N. Heat stress has immediate and persistent effects on immunity and development of Tenebrio molitor. J Insect Fungal. 2023;10(5):835–53. https://doi.org/10.1163/23524588-20230095
- 90.
Xiao B, Chen S, Wang Y, Liao X, He J, Li C. Heat Shock Factor regulation of antimicrobial peptides expression suggests a conserved defense mechanism induced by febrile temperature in arthropods. eLife. 2024. https://doi.org/10.7554/eLife.101460.1
- 91. Velásquez AC, Castroverde CDM, He SY. Plant-pathogen warfare under changing climate conditions. Curr Biol. 2018;28(10):R619–34. pmid:29787730
- 92. Vajedsamiei J, Melzner F, Raatz M, Morón Lugo SC, Pansch C. Cyclic thermal fluctuations can be burden or relief for an ectotherm depending on fluctuations’ average and amplitude. Funct Ecol. 2021;35(11):2483–96.
- 93. Bruno JF, Selig ER, Casey KS, Page CA, Willis BL, Harvell CD, et al. Thermal stress and coral cover as drivers of coral disease outbreaks. PLoS Biol. 2007;5(6):e124. pmid:17488183
- 94. Cook T, Folli M, Klinck J, Ford S, Miller J. The relationship between increasing sea-surface temperature and the northward spread ofPerkinsus marinus(Dermo) disease epizootics in oysters. Estuar, Coast Shelf Sci. 1998;46(4):587–97.
- 95. Zhang Y-B, Yang A-P, Zhang G-F, Liu W-X, Wan F-H. Effects of simulated heat waves on life history traits of a host feeding parasitoid. Insects. 2019;10(12):419. pmid:31771090
- 96.
Leicht K. Implications of heat waves on immune defence, life history traits, and adaptive potential: a snail’s perspective. University of Jyväskylä; 2014. https://urn.fi/URN:ISBN:978-951-39-5693-6
- 97. Duffy MA. Selective predation, parasitism, and trophic cascades in a bluegill-Daphnia-parasite system. Oecologia. 2007;153(2):453–60. pmid:17497181
- 98. Lian X, Huang J, Li H, He Y, Ouyang Z, Fu S, et al. Heat waves accelerate the spread of infectious diseases. Environ Res. 2023;231(Pt 2):116090. pmid:37207737
- 99.
Tompkins D, Dobson A, Arneberg P, Begon M, Cattadori I, Greenman J, et al. Parasites and host population dynamics. In: Hudson PJ, editors. The Ecology of Wildlife Diseases 2. Oxford: Oxford Academic; 2002. p. 45–62.
- 100. Frainer A, McKie BG, Amundsen P-A, Knudsen R, Lafferty KD. Parasitism and the Biodiversity-Functioning Relationship. Trends Ecol Evol. 2018;33(4):260–8. pmid:29456188
- 101. Altizer S, Ostfeld RS, Johnson PTJ, Kutz S, Harvell CD. Climate change and infectious diseases: from evidence to a predictive framework. Science. 2013;341(6145):514–9. pmid:23908230
- 102. Varotsos CA, Mazei YA. Future Temperature Extremes Will Be More Harmful: A New Critical Factor for Improved Forecasts. Int J Environ Res Public Health. 2019;16(20):4015. pmid:31635142
- 103. Harvell CD, Mitchell CE, Ward JR, Altizer S, Dobson AP, Ostfeld RS, et al. Climate warming and disease risks for terrestrial and marine biota. Science. 2002;296(5576):2158–62. pmid:12077394