2.6 Statistical analysis
One-way ANOVA was conducted to test the differences in variables between short-term and multi-year elevation of [CO2] at each observation time, and the differences in parameters measured at predawn and other time points, and the differences in values among growth periods. Two-way ANOVA was performed to investigate the effects of [CO2] increase treatments, observation time points and their interactions on PSII efficiency. Pearson correlation was applied to inquiry the relationship between parameters of PSII efficiency. Statistical analysis was completed in IBM SPSS 21.0 (IBM Crop., Armonk, NY, USA).
3. Results
Overall, compared with short-term effect (one growing season) of abrupt increase of [CO2] by 200 μmol mol-1(AI200), constant increase of [CO2] (four generations) by 200 μmol mol-1 (CI) significantly increased the maximum photochemical efficiency of PSII (Fv/Fm) at jointing stage (Table 2, Fig. 2a), while this promotive effect was not observed at heading and grain filling stages (Table 2, Fig. 2b, c). Stepwise increase of [CO2] (four generations) by 160 μmol mol-1 (SI) dramatically improved Fv/Fm than short-term (one growing season) abrupt increase of [CO2] by 160 μmol mol-1 (AI160) during three growth periods (Table 2, Fig. 2d-f). Moreover, the probability that trapped exciton moves an electron further than QA-o), electron transport quantum yield (φEo) and performance index (PIABS) under SI were all prominently higher than that under AI160 at grain filling stage (Table 2, Figs. 4f, 5f, 6f), but under CI there was only higher PIABS than under AI200 at jointing stage with no significant difference in ψo and in φEo (Table 2, Fig 6a).
The Fv/Fm, ψo, φEo and PIABS all showed significant diurnal variation, and the p -values of them among observation time points were all < 0.001 (Table 2). Fv/Fm gradually decreased to the minimum value at midday, and then increased progressively back to predawn level (Fig. 2). That was opposite to the changing trend of PPFD in a day (Fig. 1). Fv/Fm declined with the increase of PPFD (p < 0.001) and air temperature (p < 0.1) over three growth stages (Fig. 3).
The daily trend of ψo and φEo both showed the shape of letter “M”, with the two peak values both appeared at 7:00 and 17:00 at jointing and heading stages, and at 9:00 and 15:00 during grain filling stage in all treatments (Fig. 4, 5). The PIABS changed during three phases: started rising from predawn to 7:00, then decreased to minimum value at midday, and then rising again (Fig. 6). There was no obvious correlation between the diurnal change of ψo, φEo and PIABS and PPFD or air temperature (correlation coefficient was not shown).
Significant effects of the interaction between [CO2] increase treatments and observation time points on Fv/Fm, φEo and PIABS were detected, indicating that effect of SI and AI160 on Fv/Fm, φEo and PIABS were influenced by observation time in a day (Table 2). SI promoted Fv/Fm by reducing midday depression at jointing stage (Fig. 2d). The relative change (RC) of Fv/Fm was 6.5% from predawn to 11:00 under SI, lower than 8.7% under AI160, and Fv/Fm recovered to 93.0%, 95.4% and 97.2% of that predawn observation at 13:00, 15:00 and 17:00, all were higher than that under AI160 which were 91.2%, 94.3% and 96.5% (Table 3). When it came to heading stage, SI only reduced the down-regulation of Fv/Fm at midday, with the RC of Fv/Fm from predawn to 11:00 being approximately 2/3 of that in AI160 (Table 3). The relative levels of PIABS between AI160and SI fluctuated over time at jointing and heading stage, which may result in an overall insignificant difference between SI and AI160 (Fig. 6d, e). At grain filling stage, SI improved Fv/Fm, ψo, φEo and PIABS at all observation time in a day, overall increased by 1.0%, 4.6%, 5.5% and 12.2%, respectively (Fig. 2-5). No significant interaction was found between effects of AI200 and CI and observation time points (Table 2).
The largest midday depression of Fv/Fmoccurred at heading stage, reduced by 9.0%-14.1% relative to predawn value (Table 3). Fv/Fm under all treatments rebounded to 99% of predawn value until 19:00 when PPFD was 0.4 μmol m-2s-1 during jointing stage, and was fully restored to predawn level at 19:00 during heading stage when the PPFD dropped to 0 (Table 3). At grain filling stage, Fv/Fm under AI200 and AI160 measured at 17:00 showed no significant difference compared with those observed at predawn, while Fv/Fm under CI and SI were still 1% lower than the predawn values (Table 3).
There was almost no significant difference between the midday ψo and the predawn value at jointing and heading stages, with the maximum RC being 13.2% - 14.6% and 9.5% - 12.2% across all treatments, respectively (Table S1). While at grain filling stage, all ψo were dramatically greater than those measured at predawn, and the highest RC was 21.5% - 28.8% (Table S1), and the φEo in a day was obviously higher than that at predawn, except at 13:00 (Table S2). From predawn to 7:00, PIABS under AI200 and CI increased 11%, but AI200 and SI caused no marked change at jointing stage, and RC at heading and grain filling stages was enhanced, while CI and SI led to lower RC than AI200 and AI160, respectively (Table S3). The largest midday attenuation of PIABS was up to 63.4% on average, which occurred during jointing period (Table S3).
The φEo was positively correlated with ψo and Fv/Fm at jointing and heading stages under all treatments, while there was insignificant correlation with Fv/Fm at grain filling stage under AI200 and SI (Table 4). Correlation coefficient between φEo and ψo was greater than that between φEo and Fv/Fm (Table 4). Moreover, PIABS was found to be positively correlated with Fv/Fm, φEo and ψo and those three correlations were weaker than the former in sequence, except that no significant correlation was tested between PIABS and ψo under AI160 during heading stage (Table 4).
The predawn Fv/Fm at jointing stage was significantly higher than that at heading and grain filling stages, but no significant difference was observed between the latter two stages (Table 5). The predawn Fv/Fm under CI and SI were correspondingly higher than those under AI200 and AI160 during grain filling stage (Fig. 2c, f). Moreover, ψo, φEo and PIABS measured at predawn reduced dramatically with the advance of growth periods in all treatments (Table 5). The reduction of Fv/Fm, ψo, φEo and PIABS from jointing to grain filling stage were 2.6%, 19.9%, 21.9% and 51.9% under CI, all were lower than those under AI200, which were 3.3%, 21.4%, 24.1% and 55.7%, respectively (Table 5). SI caused less decrease (2.8%, 20.9%, 23.3%, 54.8%) than AI160 (3.2%, 26.3%, 28.6%, 63.4%) in Fv/Fm, ψo, φEo and PIABS, but slightly more than those under CI (Table 5).