Seasonal fluctuation in climatic conditions may contribute to the decline of insects worldwide by increasing disease prevalence and spreading diseases (see Harvey et al., 2020). Insects' immune responses are influenced by climatic factors such as temperature, humidity, and rainfall (Martin and Hillyer, 2024). Climate change-driven increase in temperature can impact insects' immune responses and infections across the spatiotemporal scale (Paul et al., 2024; Reece et al., 2017). Seasonal air temperature can alter prevalence, intensity of parasitism, and fitness costs imposed by parasites. Understanding how climate change would impact parasitism and the cost of parasitism across seasons is crucial for determining the local decline of the population across seasons. Studying variation of parasitism across seasons and how climatic local factors impact this variation provides an excellent platform to estimate how changes in seasonal climatic factors might impact population decline across seasons.

Climatic factors such as temperature, rainfall, and humidity impact various aspects of insect life history traits, including immune response (Cohen et al., 2020; da Silva et al., 2021; LoScerbo et al., 2020; Mlynarek et al., 2015). Higher temperatures can increase metabolic costs and energy required to perform basal physiological functions, thereby reducing resources available for immune responses, consequently increasing parasitism or diseases in warmer months (Khan & Rolff, 2025). For example, damselfly larvae (Coenagrion puella) exposed to heatwaves showed reduced energy reserves and lowered immunity (Tüzün and Stoks, 2021). On the other hand, higher temperatures can increase the availability of food and nutrition available which can increase resources for immune response, thereby reducing parasitism or diseases in warmer months. For example, stronger immune responses were detected under higher temperatures in mosquitoes (see Murdock et al., 2012) and sepsid flies (Gourgoulianni et al., 2023). Similarly, in Lestes forcipatus and Ischnura elegans damselflies, stronger immune responses are observed in warmer seasons (Robb and Forbes, 2005; Raczyński et al., 2022). Impact of temperature, therefore, could increase or prevalence and intensity of parasitism, the direction and extent depend on the host-parasite systems and the local climatic conditions.

Insect immune responses include cellular and humoral components, with melanisation playing a major role in humoral defence against pathogens. The prophenoloxidase (proPO) cascade drives melanin production, encapsulating pathogens and removing parasites from the insect body (see Khan and Rolff, 2025; Ilvonen et al., 2018; Siva–Jothy, 2000). Environmental factors affect individual and population immune responses (measured as PO activity or degree of melanisation) thereby modify infection rates in insects (Carter et al., 2021; Ismail et al., 2024; Scharsack and Franke, 2022). Little is, however, the impact of climate fluctuations on insect immune response and seasonal parasite prevalence, albeit of high importance for determining disease risk and seasonal decline of insects under a changing climate.

Here, we aim to understand the pattern of immune response and parasite prevalence across seasons and determine the underlying climatic drivers. We hypothesised that immune response and parasitism would vary with temperature across seasons. Specifically, 1) higher temperatures in warmer months may increase immune response and reduce parasitism if thermal conditions promote physiological activity and development; alternatively, 2) parasitism may increase in warmer months if high temperatures reduce host immune investment or favour parasite development. We tested these hypotheses using the Australian common bluetail damselfly (Ischnura heterosticta) and the endoparasite gregarine (Apicomplexa: Protozoa) as a host-parasite study system. We determined melanisation as an index of immune response and measured gregarine prevalence across seasons.

We collected damselflies from a natural population located on the Wallumattagal campus of Macquarie University, NSW, Australia. The study site is a small artificial lake with an area of approximately 895 m² and a perimeter of 212 m. The lake is permanent with stagnant water flow. We surveyed the study site every month from March 2024 to February 2025, covering the entire flight season of this species. No permits were required for damselflies collection from the site, as this species is not a protected species and the field site is not part of a national park.

We captured damselflies from the field with insect-catching nets (dimensions: 1260 mm handle, 456 mm diameter hoop, 81 cm long net bag) while walking along the edge of the water body and adjacent grasslands. For each sampling day, we covered the same study area and spent approximately 30 minutes collecting 35–40 damselflies (14 sampling days with 10 control and 30–35 experimental damselflies each day for a total n = 470). We transported the damselflies to the laboratory within five minutes of capture and placed them in a plant growth chamber (Plant growth facility, Macquarie University, NSW, Australia) for acclimation for two hours. We set the temperature at 25°C and relative humidity at 80% and the light: dark cycle was set for 16:8 h.

We quantify the immune response in damselflies using the encapsulation response assay (Nagel et al., 2011). We used a needle holder to insert a sterile 3 mm (filament insertion depth was 2 mm) long nylon monofilament (diameter 0.20 mm; treated with fine sandpaper) into the body of the experimental damselflies (Koskimäki et al., 2004; Nagel et al., 2011; Rantala et al., 2000). We insert the nylon filament into the thorax below the lateral stripe (Fig. 1b). We maintained the consistency of the length of inserted filaments throughout the experimental procedures. Control animals were not manipulated. We kept all damselflies individually in a plastic drinking cup (100 ml) covered with a cotton mesh and a wood dowel for perching. We placed the cups in the growth chamber for 24 h (studies showed melanisation reaction occurred within 24 hours after implanting nylon inserts; Galko and Krasnow, 2004; Tang, 2009), after which we recorded the survival status of control and experimental animals (dead, or alive). We removed the nylon from the experimental animals with the needle holder, and the piece of nylon was placed in ethanol (70%) in an Eppendorf tube (0.5ml). Damselflies were euthanised in -30°C and randomly selected experimental individuals (n _total = 140) from each sampling were dissected to determine the presence of gregarines.

We photographed the nylon filaments at 3.2x with an OLYMPUS SZX16 stereo microscope under standard lighting using OLYMPUS cellSens imaging software. We took the filaments' images from three different angles and used ImageJ software to calculate the amount of melanin present on the insert. We measured the melanin in the form of a greyscale value (average darkness) from the filament part, which was inserted into the thorax of damselflies. We also measured the greyscale value of the part of the insert that remained outside of the damselfly thorax to check whether all the images have similar values. In ImageJ, we converted the RGB image to 8-bit greyscale and considered 0 as pure black and 255 as pure white (the lower greyscale value indicated a higher amount of melanin present on the insert). We took the average of greyscale values from three images and subtracted it from 255 (called reverse greyscale value) to make it easier to interpret (lower greyscale value, lower melanisation) (Ferguson and Sinclair, 2017). We collected monthly maximum and minimum temperature data for 2024–2025 from the Bureau of Meteorology (BOM: http://www.bom.gov.au/climate/data/index.shtml) and calculated monthly average temperature (°C), rainfall (mm), and humidity (%) for each month that we surveyed. The average autumn, spring, and summer temperatures of this study area during our collection period were 19.83 ± 1.67°C, 19.28 ± 2.61°C, and 24.06 ± 0.48°C, respectively. Average autumn, spring, and summer rainfall levels were 132.84 ± 62.96 mm, 44.54 ± 9.79 mm, 66.84 ± 37.82 mm, respectively. Average autumn, spring, and summer relative humidity were 60.70 ± 14.15%, 53.49 ± 9.16% and 60.1 ± 5.81%, respectively.

We applied the DurgaDiff function of the Durga R package to determine mean differences of greyscale value (as a measure of melanisation) and gregarine prevalence between sexes and across seasons (Khan and McLean, 2024). This R package calculated 95% confidence intervals of the mean difference by bootstrapping 1000 times. We applied a generalized linear model (GLM) to identify the effect of climatic factors (monthly average temperature, rainfall, and humidity) on melanisation and gregarine prevalence. We fitted the GLM models with melanisation/gregarine prevalence as the response variables, and climatic factors as fixed effects. We analysed all data in R version 4.0.3 (R Core Team, 2020) using packages “lme4” (Bates et al., 2014), “performance” (Lüdecke et al., 2021), and “Durga” (Khan and McLean, 2024). We used the Durga package for estimating and plotting effect sizes (Khan and McLean, 2024). All values are estimated ± standard error. For model description, please see the supplementary information.

Overall, the nylon treatment did not differentially affect the mortality of experimental damselflies compared to control damselflies, and 94.89% of animals died after 24 hours in the growth chamber. Therefore, the observed melanisation is likely to reflect baseline immune capacity rather than short-term stress responses induced by captivity.

Melanisation is positively correlated with temperature (GLM, estimate = 1.63 ± 0.50, z = 3.25, p = 0.001; Fig. 3a), but negatively correlated with rainfall (GLM, estimate = − 0.07 ± 0.02, z = -2.67, p = 0.008; Fig. 3b) and humidity (GLM, estimate = − 0.58 ± 0.13, z = -4.35, p < 0.0001; Fig. 3c) in females. Melanisation is negatively correlated with humidity in males (GLM, estimate = − 0.31 ± 0.13, z = -2.28, p = 0.02; Fig. 3c), but not with temperature and rainfall (Table 2; Fig. 3 (a-b)).

Gregarine prevalence showed negative correlation with monthly average temperatures in females (GLM, estimate = − 0.22 ± 0.12, z = -1.81, p = 0.07; Fig. 3d), but not in males (Table 2; Fig. 3d). In females, gregarine prevalence was positively with rainfall (Table 2, Fig. 3e) and humidity (Table 2, Fig. 3f), but negatively correlated with rainfall (Table 2, Fig. 3e) and humidity (GLM, estimate = − 0.05 ± 0.02, z = -2.37, p = 0.01; Fig. 3f) in males.

We found that females and males had higher melanisation which was correlated with lower gregarine prevalence in summer compared to spring and autumn. Melanisation increased with temperature but decreased with rainfall and humidity, whereas gregarine prevalence was higher in females throughout seasons and weakly negatively correlated with temperature in females. We further found that melanisation is higher in seasons when gregarine prevalence is lower.

Overall, we showed seasonal dynamics in melanisation and gregarine prevalence, with higher melanisation and lower gregarine prevalence in the warmer season. These findings highlight that seasonal climatic fluctuations shape host-parasite interactions, providing valuable insights into patterns of disease dynamics and insect fitness.