Short gaps between heatwaves hit subtropical tree seedlings hardest

When two heatwaves strike a young tree only a week apart, the damage looks very different from what appears when the same two events are separated by three weeks. That is the central finding of a new greenhouse experiment with two subtropical broadleaved species, Schima superba and Phoebe bournei, published in the journal Ecological Processes. Under the shortest inter-heatwave interval tested, net photosynthetic rate fell by 50.8 percent in S. superba and by 32.4 percent in P. bournei relative to unheated control seedlings. Photosynthetic rates were higher at longer intervals in both species.

The study was carried out at the Fujian Academy of Forestry Science in Fujian Province, China, using two-year-old seedlings grown in pots. Five treatment conditions were compared: an unheated control, a single three-day heatwave, and three recurrent-heatwave conditions in which two three-day heatwaves were separated by intervals of 7, 15, or 23 days. The first heatwave occurred on different dates among the recurrent treatments, but the second heatwave and the single-heatwave event were synchronized, and all physiological measurements were taken on a common final date, August 20, 2024. This design compared different treatment groups at one endpoint rather than tracking the recovery of the same seedlings over time.

A crucial structural limitation shapes every conclusion in the paper. Each treatment condition was implemented in only one semi-open greenhouse, and the six measured seedlings of each species within that greenhouse were biological subsamples, not independent treatment replicates. Treatment condition is therefore confounded with greenhouse identity, and greenhouse-to-greenhouse variation cannot be independently estimated. The authors state plainly that seedling-level comparisons describe variation among measured plants in these particular greenhouses and cannot provide valid replicated tests of heatwave-treatment effects. All treatment comparisons and exploratory statistics in the study must be read with this restriction in mind.

The heating itself was simulated with semi-open transparent greenhouses equipped with 3500-watt industrial heaters suspended about 1.5 meters above the ground at a 70-degree angle to reduce direct exposure of seedlings to heated airflow. During heated periods, daytime air temperature differences relative to the control were approximately 5 to 10 degrees Celsius, nighttime differences 0 to 5 degrees, with a mean reported air temperature difference of about 7.1 degrees Celsius. Soil water content in the heated greenhouses was roughly 6 percent lower on average than in the control greenhouse. The authors note that the treatments thus differed in both temperature and evaporative and soil-moisture conditions, and that these co-occurring differences should be considered when interpreting the physiological responses.

On the final sampling day, the researchers measured leaf gas exchange with a LI-6800 portable photosynthesis system, recording net photosynthetic rate, transpiration rate, stomatal conductance, and instantaneous water-use efficiency on mature, fully expanded sun-exposed leaves. Chlorophyll a fluorescence was measured with a Pocket-PEA fluorometer after 30 minutes of dark adaptation, yielding the maximum photochemical efficiency of photosystem II, Fv/Fm, the performance index PIABS, and a set of OJIP-derived energy fluxes per active reaction center and per excited cross section. Leaf samples were also collected to determine the activities of the antioxidant enzymes superoxide dismutase, peroxidase, and catalase, and the content of malondialdehyde, an indicator of membrane lipid peroxidation, using standard spectrophotometric assays on a fresh-weight basis.

The gas exchange results showed a clear interval pattern. In both species, net photosynthetic rate increased progressively from the 7-day treatment to the 23-day treatment, approaching the levels observed after a single heatwave. Under the 7-day interval, transpiration and stomatal conductance were approximately 32 and 36 percent lower than control in S. superba, and 37 and 46 percent lower in P. bournei. Transpiration rose with longer intervals in both species. Water-use efficiency, however, split the two species: it was higher at 23 than at 7 days in S. superba, shifting from negative to positive carry-over effect sizes, whereas P. bournei showed the opposite pattern, with a slightly positive effect size under the 7-day interval that turned negative at 15 and 23 days. The nominal species-by-interval interaction for water-use efficiency was significant at P < 0.001 in the exploratory analysis.

Oxidative responses also diverged between the species. Malondialdehyde content peaked under the 7-day interval in S. superba, indicating greater lipid peroxidation in that treatment group, but peaked under the 15-day interval in P. bournei, where its lowest level occurred under the 23-day interval. Catalase activity in S. superba was lower under the 7-day interval than after a single heatwave but increased under the 15-day interval, while peroxidase activity declined most markedly under the 15-day interval. In P. bournei, catalase remained relatively low under the 7-day interval and rose with longer intervals toward the control level, superoxide dismutase generally declined from the 7-day to the 23-day interval, and peroxidase stayed comparatively stable. Exploratory analyses found significant nominal interval associations for all four indicators, and significant species-by-interval interactions for catalase, peroxidase, and malondialdehyde, but not for superoxide dismutase.

Chlorophyll fluorescence told a partly different story. Fv/Fm reached its lowest value under the 7-day interval in both species; in S. superba it rose with longer intervals to a level comparable to the control under the 23-day interval, while in P. bournei it was higher under the 15-day interval than under the 7-day interval but decreased slightly under the 23-day interval. In S. superba, the energy flux parameters ABS/RC, DIo/RC, and TRo/RC were relatively high under the 7-day interval and decreased at longer intervals, whereas P. bournei maintained comparatively high values under both the 7-day and 23-day intervals with lower values under the 15-day interval. Notably, ETo/RC under the 7-day interval in S. superba was slightly higher than under the control and single-heatwave treatments rather than showing a reduction, showing that individual fluorescence parameters did not move uniformly. No significant species-by-interval interactions were detected for the fluorescence parameters.

The association analyses added a further layer of species contrast. In S. superba, catalase was the only antioxidant indicator significantly associated with photosynthetic rate, and the exploratory path model showed a negative catalase-photosynthesis association with a standardized path coefficient of -0.301, while Fv/Fm showed a positive association with a coefficient of 0.265. In P. bournei, peroxidase was strongly associated with photosynthetic rate, with a positive path coefficient of 0.287, and was also significantly associated with transpiration, stomatal conductance, and water-use efficiency. Stomatal conductance showed significant positive associations with photosynthetic rate in both species, with coefficients of 0.709 and 0.686 respectively. Random forest analysis identified stomatal conductance and photosynthetic rate as important predictors of peroxidase activity in S. superba, and Fv/Fm and photosynthetic rate as important predictors in P. bournei, while chlorophyll fluorescence parameters contributed 83 and 60 percent of the grouped predictor importance for malondialdehyde in the two species.

The authors are careful about what these patterns do and do not show. Because measurements were taken at a common endpoint, the observed differences represent interval-associated physiological responses rather than directly measured recovery trajectories. The exploratory exponential curves fitted to the interval-specific effect sizes had no residual degrees of freedom with only three intervals, so they are retained solely as visualizations, not as validated recovery kinetics. The modeled pathways are statistical associations rather than confirmed causal mechanisms, and the lack of independent greenhouse replication and repeated physiological measurements limits treatment-level inference and prevents direct characterization of recovery kinetics. The team calls for future experiments with independent replication, repeated measurements across broader interval ranges, and direct measurements of leaf temperature, reactive oxygen species, and photosystem repair. Within those limits, the study argues that evaluating heatwave carry-over effects in trees requires integrating carbon-water relations, oxidative stress, and photochemical performance together, because the response of any single trait can mislead.

Subject of Research: Physiological carry-over effects of inter-heatwave interval length on photosynthesis, oxidative stress, and chlorophyll fluorescence in two subtropical tree seedling species

Article Title: Physiological carry-over responses across inter-heatwave intervals in two subtropical tree species

Article References: Qu, L.-P., Wang, S., Lin, J., Tang, X., Deng, Z., Zhang, R., Liu, Y., Yan, X.-L., Shao, C., & Hu, Y. (2026). Physiological carry-over responses across inter-heatwave intervals in two subtropical tree species. Ecological Processes, 15(1), Article 101. https://doi.org/10.1186/s13717-026-00760-y

Image Credits: AI Generated

DOI: 10.1186/s13717-026-00760-y

Keywords: heatwaves, carry-over effects, Schima superba, Phoebe bournei, photosynthesis, water-use efficiency, chlorophyll fluorescence, antioxidant enzymes, malondialdehyde, subtropical trees, seedlings, inter-heatwave interval