*3.4. Impact of Drought on Ecosystem Carbon Fluxes at Monthly Timescale*

The monthly averages of daily sums of GPP and Reco were analysed separately to assess whether the observed net ecosystem productivity (NEP) reduction during drought was mainly a consequence of increased Reco, reduced GPP, or contribution of both across the forest ecosystems (Figures 5–7).

**Figure 3.** Relationship between the light response curve (LRC) residuals of the light response curve of gross primary production and the vapour pressure deficit (VPD) at the spruce forest sites in CZ-BK1 and CZ-RAJ for the normal (NY) and dry years (DY). Red represents low SVWC conditions and blue shows high SVWC conditions. The grey dashed lines represent the piecewise regression model slope, whereas the black dashed lines show the breakpoint values.

**Figure 4.** Relationship between the light response curve (LRC) residuals of the light response curve of gross primary production and the soil volumetric water content (SVWC) at the spruce forest sites in CZ-BK1 and CZ-RAJ for the normal (NY) and dry years (DY). Red represents high VPD conditions and blue shows low VPD conditions. The grey dashed lines represent the piecewise regression model slope, whereas the black dashed lines show the breakpoint values.

**Table 7.** Breakpoints and Slopes from the piecewise regression of the light response curve residuals to vapour pressure deficit (VPD) and soil volumetric water content (SVWC) for the period of May–September.


Signif. code for the breakpoint values: *p* < 0.001 '∗∗∗'.

**Figure 5.** Monthly averages of daily sums of gross primary productivity (GPP) for May–September of the normal years (NY) and dry year (DY) in CZ-BK1 and CZ-RAJ. The tables within the figure represent the mean monthly vapour pressure deficit (VPD) and soil volumetric water content (SVWC) values from May to September for each forest station.

**Figure 6.** Monthly averages of daily sums of ecosystem respiration (Reco) for May–September of the normal years (NY) and dry year (DY) in CZ-BK1 and CZ-RAJ. The tables within the figure represent the mean monthly air temperature (Tair) and soil volumetric water content (SVWC) values from May to September for each forest station.

**Figure 7.** Monthly averages of daily sums of NEP for May–September of the normal years (NY) and dry year (DY) in CZ-BK1 and CZ-RAJ.

It was observed that there was a significant decline in the total GPP by 14% and 6% during the main growing season period of the dry year for CZ-RAJ and CZ-BK1, respectively. There were also observed significant statistical differences (*p* < 0.001) in the mean monthly GPP values recorded for July and August during the dry year as compared

to the years characterized by normal conditions at both forest sites, especially for CZ-RAJ (Figure 5). Moreover, this significant decline in the mean monthly GPP values at both spruce forest stands during the months of July and August coincided with high daily mean VPD values (>12 hPa) and low SVWC values (<0.16 m<sup>3</sup> m−3) as compared to the same period for the adjacent years with normal conditions.

However, an increase in the monthly mean Reco values was observed for July to September of the dry year as compared to the adjacent normal years within CZ-BK1 but with no significant statistical differences, except for the month of August (Figure 6). These three months within 2015 were also characterized by high mean Tair (>20 ◦C) and low mean SVWC (<0.19 m3 m−3) as compared to the two other adjacent years. In addition, during the DY period within CZ-RAJ, the monthly mean Reco values were found to have declined significantly, as compared to the adjacent NY periods. We also observed a 38% significant decline in the mean monthly NEP values during the dry year period for the dry spruce forest in Rájec, compared to a 12% decrease in the mean monthly NEP values for the humid spruce forest in Bílý Kˇríž (Figure 7). There were also statistically significant differences (*p* < 0.001) in the mean monthly NEP for the months of July and August across both spruce forest stations. During these months, the observed mean monthly NEP values largely declined during the DY period, as compared to the same period during the NY, especially in CZ-RAJ.

To summarize, the large decline in GPP, Reco and NEP during the dry year period (especially from July to August) for CZ-RAJ showed that the impact of the drought was more severe in CZ-RAJ than in CZ-BK1. However, during July–August of the DY period, Reco at CZ-BK1 significantly increased as compared to that in NY.

### **4. Discussion**

The study sought to assess the different effects of the 2015 summer drought on wet and dry spruce forest ecosystems. Our findings corroborate the earlier published results by [27–29] that the year 2015 was characterized by severe drought conditions across Europe (Figure 1 and Table 4). Furthermore, it was found that the months of July–August of 2015 were the most affected by high mean Tair (>20 ◦C), high mean VPD (>10 hPa) and low mean SVWC values (<0.19 m<sup>3</sup> m<sup>−</sup>3) across both spruce forest stands. This shows that the drought effect in 2015 was severe in the months of July–August. Moreover, the results of this study also show that the rate of the forest ecosystem photosynthesis was significantly reduced (likely through the immediate closure of the stomata to protect the tree from desiccation) by the high VPD, Tair and soil water deficit during these two months in 2015. As such, there was an observed strong decline in both forest GPP and NEP across both spruce forest stands during the DY period of 2015, especially in CZ-RAJ compared to CZ-BK1 (M). This strong decline in forest GPP and NEP at CZ-RAJ was mainly due to the high atmospheric evaporative demand coupled with the low SVWC conditions experienced at this forest ecosystem, as compared to that in CZ-BK1 [30]. However, due to the humid climatic conditions at CZ-BK1, an increase in Tair during the months of July–August in 2015 only aided the increase in the kinetics of the enzymes participating in microbial decomposition and root respiration under warm conditions, thereby increasing the overall forest ecosystem respiration [65–71].

Additionally, results from the piecewise regression analysis using the residuals from the LRC model highlight the effects of both VPD and SVWC on forest ecosystem GPP. This is because both photosynthesis and transpiration are mediated by stomatal conductance, which are affected by these environmental variables. However, a steeper decline in forest ecosystem GPP (Figure 3 and Table 7) was observed with high VPD values across both forest stands, even under non-drought years (when SVWC was non-limiting). This is consistent with recent studies, which show that high VPD values aggravate drought effects in forests, due to the abrupt changes over very short timescales within a day, even without dry soil conditions [7,72,73]. This further explains the strong suppression of GPP by high VPD on even non-drought years (2014 and 2016), as the SPEI (for determining dryness)

only captures changes in the contributing environmental drivers (temperature and soil moisture) for longer time scales of weeks or months [74,75]. Thus, we would recommend to include the influence of these abrupt changes in VPD on the rate of photosynthesis with SVWC limitations in future LRC models when analysing the impact of drought on GPP, especially for different forest ecosystems that are exposed to regular strong edaphic droughts [49,62,76,77].

#### **5. Conclusions**

The study shows that atmospheric constraints increase the vulnerability of the Norway spruce forest to the severity of drought, especially at sites with a moderately dry climate that are characterized by precipitation that is typically equal or smaller than the atmospheric evaporative demand. It also further highlighted the strong influence of VPD on carbon uptake, which was further worsened by the decline in soil moisture. The effect of SVWC on GPP was especially noticeable during severe drought conditions within the DY period. Consequently, with regards to climate change, our results suggest that elevated temperatures will further exacerbate the drought impacts on forest (Norway spruce) ecosystems at sites with precipitation levels equal or smaller than the atmospheric evaporative demand, such as CZ-RAJ. The decline in GPP and NEP in 2015 found in our study thus questions not only the sustainable productivity, but also the existence of Norway spruce per se in such areas, considering the prolonged period of drought in future climatic conditions. The results of this study may help decision makers to quantitatively assess the performance of Norway spruce in future climatic conditions.

**Author Contributions:** Conceptualisation, C.M. and L.Š.; methodology, C.M., L.Š., M.F., M.P. and M.V.M.; software, C.M., L.K. and L.Š.; validation, M.F., M.A., G.J., J.D.M. and N.K., E.K.N.; formal analysis, C.M. and L.Š.; investigation, C.M. and L.Š.; resources, M.P. and L.Š.; data curation, L.Š. and G.J.; writing—original draft preparation, C.M.; writing—review and editing, L.Š., M.F., M.A., G.J., L.F., J.D.M., N.K., E.K.N. and M.V.M.; visualisation, L.K.; supervision, L.Š. and M.V.M.; project administration, M.P. and L.Š.; funding acquisition, M.P. and M.V.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research and the APC was funded by the Ministry of Education, Youth and Sports of the Czech Republic (CR) within the CzeCOS program, grant number LM2018123.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on FLUXNET and could also be made available on request from the corresponding author.

**Acknowledgments:** The authors wish to thank Radek Czerný who assisted in the eddy covariance data processing. This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic (CR) within the CzeCOS program, grant number LM2018123. L.Š. was supported by the Ministry of Education, Youth and Sports of CR within Mobility CzechGlobe2 (CZ.02.2.69/0.0/0.0/18 053/0016924). M.F., L.Š. and L.F. acknowledges the support by the Ministry of Education, Youth and Sports of the Czech Republic for SustES—Adaptation strategies for sustainable ecosystem services and food security under adverse environmental conditions (CZ.02.1.01/0.0/0.0/16 019/0000797). M.V.M. acknowledges the support by VEGA project 2/0013/17 (Slovak Agency for Scientific Support): The role of ecosystem services in support of landscape conservation under the global change.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
