Impact of temperature and hypoxia on the size and survival of aquatic insects

It is crucial to understand impact of the globally rising temperatures on the functional traits of the insects. This is mostly due to the functional traits change, under temperature regime shift, can impact survivability of the individuals and population. Insects´ size is an important functional trait, affected by temperature. It was hypothesized, that many aquatic insects are getting smaller, with the temperature increase, probably due to the hypoxic-induced reduction of metabolic rate.

This dataset contains data on the larvae and pupae size of non-biting midge (Chironomus riparius) and their change in the mesocosm setting, and their change under the impact of different combinations of temperature and hypoxia.

We have compared the body size and survivability of C. riparius in six experimental treatments: three at 20°C, with high, medium and low oxygen concentration and three at 30°C, with the same three levels of oxygen concentrations. To achieve this comparison, we have collected and measured pupal exuviae and (at the very end of the experiment) larvae of the non-biting midge.

This dataset contains digital measutments of the scaled images of larvae and pupal exuviae. All the exuviae photos are available at https://doi.org/10.5281/zenodo.14516968. Measurements were conducted on the digital images, using Fiji –ImageJ with calibrated scale (scales for the calibration are attached alongside the rest of the images at Zenodo)

Our measurments https://zenodo.org/uploads/14516968 showed that C. riparius pupae were significantly smaller at 30°C treatments with medium and low oxygen levels than in the rest of the treatments. No significant reduction in pupal size was observed in the 30°C treatment with high oxygen content. We have found no significant differences in size of larvae between the treatments. However, we detected reduced larval survivability in the 30°C treatments with medium and low oxygen levels compared to the other treatments.

Methods
Experimental design

Our experimental system utilized laboratory-reared populations of the non-biting midge Chironomus riparius (Meigen, 1804). The mesocosm experiment was conducted from November 10, 2023, to March 23, 2024. Each experimental treatment included four replicate units, with each replicate consisting of a 2.7 L bucket placed inside an insect rearing cage. These cages were housed in a climate-controlled chamber maintained at 20 ± 2 °C, 60 % relative humidity, and an 8:16 hour light:dark cycle. Each bucket was filled with 100 g of clean sand and 2.7 L of dechlorinated water. All replicates for a given treatment were placed within the same cage and climate chamber. Although larvae and pupae developed in isolated water-filled buckets with minimal inter-bucket interaction, replicates within the same cage must be regarded as pseudoreplicates.

We established six treatment conditions: Treatments A, B, and C ("Lower" temperature group) were maintained at the chamber's background temperature, ranging from 18.9 to 22 °C. Treatments D, E, and F ("Warmer" group) were exposed to elevated temperatures via submersible aquarium heaters set to 30 °C, resulting in actual temperatures between 26.8 and 28.5 °C.

Specimens collection

Pupal exuviae of emerging C. riparius adults were collected daily throughout the duration of the experiment. We focused exclusively on collecting exuviae to avoid disturbing the C. riparius populations within the experimental setups; no adult specimens were captured. At the conclusion of the experiment on March 23, 2024, all remaining live C. riparius larvae were collected from each treatment. No larval sampling was conducted during the experimental period in order to preserve the integrity of the developing populations. Pupal exuviae and larvae were preserved in 70 % ethanol and, within several weeks of collection, mounted on permanent slides using Hydromatrix™ mounting medium. Both life stages were mounted in a dorsal-ventral orientation.

Imaging

After the slides had dried, each specimen was photographed using a Leica Ivesta3 stereomicroscope equipped with a Canon EOS 2000D camera. The resulting images were processed and analyzed in ImageJ software (version 1.54j) (Schindelin et al., 2012). Calibration and measurement of the images were performed in ImageJ using object micrometers photographed at the same magnification as the specimens. For larval specimens, head capsule length was measured from the posterior margin of the occipital sclerite to the ventral edge of the submentum (see Fig. 1A in associated paper). In pupal exuviae, total body length was measured from the anterior margin of the frontal apotome to the posterior edge of the anal lobes (see Fig. 1B in associated paper).

Data Analysis

All statistical analyses were performed using R version 4.2.3 (2023-03-15 ucrt) — "Shortstop Beagle". To assess within-group differences in body size across treatments, we applied one-way ANOVA followed by Tukey's post hoc test, as the size data met the assumptions of normality.

To further explore pupal size variation, we developed two generalized linear mixed models (GLMMs) using a log-link function, implemented in the glmmTMB package. The first model included an interaction term between the two primary predictors: oxygen saturation and temperature:

[1]
glmmTMB(Length of pupae ~ Temperature * Oxygen saturation + Date + (1 | Replicate bucket) + ar1(Date + 0 | Replicate bucket), family = Gamma(link = "log"))

In this model:

Length of pupae denotes the measured total length of the pupal exuviae.

Temperature represents the average experimental temperature per treatment.

Oxygen saturation is the mean oxygen saturation for each treatment.

The Temperature * Oxygen saturation term captures the interaction between these two predictors.

Date refers to the date of exuviae sampling.

Replicate bucket indicates the individual pseudoreplicate container within a treatment, specified as a random effect.

An autoregressive (AR1) correlation structure was included for Date within each replicate bucket to account for temporal autocorrelation due to repeated measurements over time.

A second, simplified model was also constructed, excluding the interaction term and including only the main effects:

[2]
glmmTMB(Length of pupae ~ Temperature + Oxygen saturation + Date + (1 | Replicate bucket) + ar1(Date + 0 | Replicate bucket), family = Gamma(link = "log"))

References

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., Saalfeld, S., Schmid, B., Tinevez, J. Y., White, D. J., Hartenstein, V., Eliceiri, K., Tomancak, P. & Cardona, A. (2012). Fiji: an open-source platform for biological-image analysis. Nature Methods, 9(7), 676-682. https://doi.org/10.1038/nmeth.2019