*3.2. Light Response Curves of GPP at Different Climates*

The parameters of LRC for normal years (NY) 2014 and 2016 were compared with those for DY for both experimental forest sites (Figure 2; Table 6). Generally, the apparent quantum yield (*α*) at CZ-BK1 was found to be higher than that in CZ-RAJ. During the years with normal conditions, *α* in CZ-BK1 was observed to be 12% higher than in CZ-RAJ when PAR was less than 500 μmol m−<sup>2</sup> s<sup>−</sup>1. Moreover, during the DY of 2015, *α* at CZ-RAJ further declined by 25% as compared to that in CZ-BK1.

The maximum gross primary production at light saturation (GPPmax) for CZ-BK1 was found to be higher than that for CZ-RAJ during the entire study period. During the NY period, GPPmax in CZ-BK1 was 34% higher than in CZ-RAJ. However, during the dry year period, there were significant reductions in the GPPmax values recorded at both spruce forest stations with (18% decline in CZ-BK1 and 17% decline in CZ-RAJ) as compared to those recorded for the years with normal conditions. Additionally, it was further observed that the GPPmax value recorded at CZ-RAJ during the DY period was 33%

lower than that in CZ-BK1. Interestingly, the GPPmax value recorded in CZ-BK1 during the DY period was still higher than the GPPmax value recorded at CZ-RAJ during the years with normal conditions.

**Figure 1.** Histogram showing the frequency of occurrence of (**a**) vapour pressure deficit (VPD) and (**b**) soil volumetric water content (SVWC) conditions during May–September between years with normal conditions (dark colour) and drought stress (grey colour) in CZ-BK1 and CZ-RAJ.

**Table 6.** Light response curve parameters for the normal (2014 and 2016) and dry (2015) years for the wet and dry climates (CZ-BK1 and CZ-RAJ respectively) within the main growing season period of May–September. The apparent quantum yield (*α*), the maximum gross primary production at light saturation (GPPmax) and the coefficient of determination (R2) are also shown.


**Figure 2.** Response of gross primary productions to photosynthetically active radiation during the years with normal conditions (black) and affected by drought stress (red) in CZ-BK1 and CZ-RAJ. The half-hourly GPP values (points) were fitted using the logistic sigmoid light response curves (lines).

#### *3.3. Response of Light Response Curve Residuals to VPD and SVWC at Different Climates*

The applied LRC model reflects only the light response of GPP and thus, its residuals can be used to assess the influence of additional meteorological parameters (i.e., VPD and SVWC) that are known to affect GPP (Figures 3 and 4). The residual analysis confirmed that the residuals consistently tended to be more negative with increasing VPD and decreasing SVWC (dry conditions). VPD had a significant and stronger effect on GPP than SVWC across both sites, as seen from the piecewise regression analysis, using the residuals from the LRC for the years under study. However, during the DY period, both VPD and SVWC had significant effects on GPP at CZ-RAJ. For CZ-BK1, there was a minimal effect of SVWC on GPP during the years with normal conditions.

Generally, the relationship between the residuals and changes in VPD and SVWC revealed a biphasic response to drought, except for the DY period in CZ-RAJ in Figure 3. All the breakpoints from the piecewise regression were found to be statistically significant (Table 7). However, steeper slopes after the VPD breakpoint values from the initial slope were mostly observed for both DY and NY periods in CZ-RAJ and only for DY period in CZ-BK1. This shows that during all the years under study, GPP decreased at a faster rate after the breakpoint in CZ-RAJ than in CZ-BK1, except for the DY period when there was a significant decline in GPP at a faster rate after the breakpoint across both sites.
