the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Long-term nitrogen fertilization increases drought sensitivity of gross primary productivity capacity in a boreal Scots pine forest
Abstract. Nitrogen (N) is a key limiting element for plant photosynthesis in boreal forests. Thus, N fertilization is proposed as an effective management strategy to increase forest productivity and the associated carbon (C) sink in the N-limited boreal biome. However, there is a limited understanding of how N fertilization can affect the sensitivity of the C sink to drought stress, which is predicted to occur more frequently in the boreal region in a changing climate. This study was based on a 15-year controlled N fertilization experiment in a boreal Scots pine stand. Ecosystem light-saturated photosynthetic capacity (GPP2000) is a good indicator of forest photosynthesis response to environmental stress. Here, we used eddy covariance measurements of C fluxes data and environmental data from paired sites to investigate whether long-term N fertilization altered the drought sensitivity of the GPP2000. We found that long-term N fertilization significantly increased ecosystem GPP2000 even on dry days during summer (June, July, and August). However, a significantly divergent drought sensitivity of GPP2000 between the N Fertilized and Reference sites was detected. Specifically, N fertilization increased the sensitivity of GPP2000 to both atmospheric and soil drought to the extent that it may offset the positive effect of N fertilization on GPP2000. Moreover, using the random forest model, we found that the absolute GPP2000 difference between fertilization and control sites was mainly influenced by air and soil drought proxies followed by canopy conductance rather than the air temperature. These results advance our understanding of the mechanisms of forest response to drought with long-term N fertilization.
- Preprint
(1612 KB) - Metadata XML
-
Supplement
(505 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2026-1940', Anonymous Referee #1, 17 Jun 2026
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2026/egusphere-2026-1940/egusphere-2026-1940-RC1-supplement.pdfCitation: https://doi.org/
10.5194/egusphere-2026-1940-RC1 -
RC2: 'Comment on egusphere-2026-1940', Anonymous Referee #2, 27 Jul 2026
Chen and others investigate the impacts of fertilization and drought on canopy carbon dioxide uptake in Scots pine. Results suggest that the fertilized forest is more sensitive to drought, consistent with the notion that they become hydrologically ‘top-heavy’ if investments in root area index are rendered less important by ample N and overshot by investments in leaf area index. The analysis makes some interesting points and has strong features but the RF analysis did not sufficiently account for multicolinearity and a few other technical issues need to be addressed.28: adding some nuance from the classic Bonan studies on forest impacts of the climate system would be forthcoming.39: the original reference for this is probably Körner 1995. There are also a few topics muddled together in this paragraphEquation 1 is a good idea given that the typical rectangular hyperbola will have a difficult time finding the saturation value; the ‘bendy straw’ of the non-rectangular hyperbola helps.122: was a value forgotten here?Was the theta parameter constrained in any way to take logical values (between 0 and 1 if I’m not mistaken)? This might not be an issue but this model can be hard to parameterizeEq 7: was standardization necessary here? Many of these values are not normally distributed and the linear regression model will fit regardless of normalization; the parameters will just have units.In Figure 2, there was some missing data pre-2015 for the reference site, which is fine but were these periods excluded in the bar graph with the average statistics when comparing?Were the eddy covariance systems calibrated against each other?For the attribution analysis how was covariance amongst drivers accounted for? WAI and VPD will covary quite strongly and a standard RF implementation will overestimate the importance of the more important variable (often called a ‘greedy’ split or similar). This multicollinearity must be accounted for.Citation: https://doi.org/
10.5194/egusphere-2026-1940-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 321 | 51 | 26 | 398 | 30 | 24 | 24 |
- HTML: 321
- PDF: 51
- XML: 26
- Total: 398
- Supplement: 30
- BibTeX: 24
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1