the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Long-term forest-line dynamics in the French Pyrenees: an accelerating upward shift related to forest context, global warming and pastoral abandonment
Abstract. Worldwide, the upper forest line has climbed over the past decades, shaping mountain landscapes in response to global changes. In European mountains, this recent trend is a continuation of the forest transition initiated in the mid-19th century, when forest extent was minimal. This study aimed to reconstruct the forest-line dynamics for the entire French Pyrenees from the mid-19th century until today. To ascertain the forest-line elevational shift for the 114 municipalities studied, three digital land-use maps (dated 1851, 1993 and 2010) were employed. The forest-line shift velocity was calculated for two periods delineated by these maps. We applied linear mixed-effect models to investigate the influence of human and environmental drivers on the forest-line shift. The mean upward shift was 0.9 m.yr-1 during the 1851–1993 period but was four-fold higher during the 1993–2010 period (3.5 m.yr-1). During the first period, the forest-line shift coincided with the isotherm upward shift, resulting from global warming. However, during the second period, despite an acceleration, the forest line lagged behind the isotherm upward shift, deepening its climatic debt. Furthermore, during the first period, the forest line shifted upward seven times faster in the eastern Pyrenees, where the mountain pine, a pioneer species, formed the forest line and pastoral abandonment occurred earlier, than in the western Pyrenees (1.3 vs. 0.2 m.yr-1). Conversely, in the following period, the shift occurred three times as fast in the western Pyrenees, where abandonment became widespread, as in the eastern Pyrenees (5.6 vs. 2.1 m.yr-1). In addition, during the second period, the closed forest line climbed twice as fast as the forest line (5.6 m.yr-1), indicating a pronounced densification of the subalpine forest. Our innovative approach integrates a large spatial scale and temporal depth and sheds new light on the interrelationships between global warming, pastoral abandonment and the forest-line upward shift.
- Preprint
(1586 KB) - Metadata XML
-
Supplement
(2577 KB) - BibTeX
- EndNote
Status: open (extended)
-
RC1: 'Comment on egusphere-2024-4099', Anonymous Referee #1, 07 Feb 2025
reply
The manuscript assesses the changes in forest-line dynamics in the French Pyrenees between 1850 and 2010. Although changes in forest surface and treeline position have been extensively studied in most European mountain ranges, this study covers an impressive temporal and spatial range, which makes it very valuable. Moreover, the methodology developed to assess changes in forest line position is original and adequate.
I do have, however, some concerns and comments I would like to transfer to the authors of the study, and that I organized by sections:
INTRODUCTION
I find the introduction a bit disorganized, although the information included in it is valuable and complete. For example, you explain the dynamics of forest line position and some of the potential drivers, then provide examples of studies reporting changes in Europe, to then explain the potential of historic maps and then move back to the effects of climate on forest line position. As I said, you probably don’t need to rewrite or delete anything, but try to build a better flow of ideas, from the generalities of treeline position to the particular case of the Pyrenees and then try to detect the gap of knowledge that your study covers, and how the use of historical maps can help you in the process.
I also wondered why you chose the data sets. I understand that the Napoleonic map is probably the first record of forest cover available, and it is indeed a very valuable product. But then why wait until 1993? Aren’t there any other forest cover maps available? Accordingly, your last map dates t0 2010, i.e. 15 years ago. Aren’t there newer editions of the forest map?
METHODOLOGY:
You repeat many times the structure “To account for…, we calculated”, or “To characterize…, we calculated”. This kind of breaks the reading flow. You can use “We calculated… to characterize…” sometimes (or alternative phrasings), it will help the readability of the text.
To estimate the distribution of forest cover according to elevation you define 100 m elevational bands. Why did you choose 100 m? This seems a bit wide considering the typical change rates in forest expansion. Did you try with different band widths to assess if the results changed?
If I understood it right (this part was a bit confusing) climate change was assessed based on changes in climate at the Pic-du-Midi station, ignoring any potential spatial variation in change. This is hard to justify, and I’d encourage the authors to gather spatially explicit climate data that allow them to assess the spatial variations in climate change, I am sure they may exist for the Pyrenees.
Can you assess which part of the livestock grazed in the mountains? Some of the animals included in the number (e.g. pigs) are typically stabled and are very unlikely to produce any effect on forest expansion. Moreover, one of the most significant changes in the economic structure of the Pyrenees, at least on the Spanish side, has been the replacement of extensive livestock (mainly sheep) by intensive livestock farming, and this is not captured by your data. With all this, it is not surprising that livestock dynamics are so closely related to population density, as they are probably capturing changes in population numbers rather than in farming patterns.
RESULTS, MAPS, TABLES AND FIGURES
Figure 1: the visual quality of the maps could be improved. For example, provide some background (DEM, hillshade) for the territory out of the study area, or the location of the main cities or villages, etc. Panel A is also missing scale and background
Maybe it is only in the pdf generated by the journal, but the resolution of figures is really low. In Figure 2 and 3, for example, the lines appear choppy and pixelated.
Table 1 contains a lot of information. I suggest keeping mean and sd and maybe move the range to Supplementary Materials.
Figure 4. The figure caption is not too clear in my opinion. For instance, it is not clear what the color scale in the right panels is actually showing.
Figure 5. Please modify the x axes so that it is clearer than “Closed forest line” only refers to the last three points. Probably it would be clearer if the “closed forest” data was represented in a different panel. Moreover, I would suggest using boxplots and/or violin plots to better reflect the distribution of values.
Figure 6. Please see my comments for Figure 1. Avoid showing the polygons as “floating” in a white space. And please use a colorblind-friendly palette.
DISCUSSION
You found a much higher rate of change than previous studies, some in the same study area. Most of the expansion, though, occurred in the first period, between 1850 and 1993. Couldn’t this be related to potential deficiencies or differences in the Napoleonic map compared to the other two products? This possibility should at least be more thoroughly discussed.
SPECIFIC COMMENTS
L36: please indicate in which regions these palynological studies have been conducted
L41-44: these two sentences seem to contradict each other.
L49-50: palynological and dendrochronological studies indeed cover small areas, but I question this is the case with historical maps.
L111: “separating France to the north from Spain to the south,” this sentence reads a bit awkward
L116-118. Are these values for the French Pyrenees or for the whole range? Please clarify and provide values for the study area (French Pyrenees) if possible.
L143-144. In those municipalities in the border with Spain, the 1 km buffer must enter into Spanish (or Andorra) territory. Did you add information from the forest cover in those two countries? How did you deal with those cases?
L167: “for each municipality (…) we rasterized the forest in each vector dataset”. This sentence is unclear, please rephrase
L222: this new data source appears here for the first time. Why? Please explain clearly from the beginning that four data sources will be used.
L247: if you calculate change in livestock density between 1852 and 2000, then why gathering the data from 1838 (line 243)?
L255: please provide more information on how the models were built
L284: It is highly standard to use ΔAIC < 2 to identify better performance models. I know this is a convention and can be changed, but I am curious to know the reason to set this threshold to 1.75
L355: “a lower mean summer water balance”. This is unclear, do you mean drier?
L400-408: the position of the treeline is typically not determined by the mean annual temperature but why some combination of temperatures during the growing season and the length of the growing season (see papers by Christian Körner). Do you know if the trend in temperatures in the study area was similar for annual temperatures than for those related to the position of timberline? Else, you may be under- or overestimating the expected change in potential forest line position.
Citation: https://doi.org/10.5194/egusphere-2024-4099-RC1
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
109 | 30 | 5 | 144 | 14 | 1 | 3 |
- HTML: 109
- PDF: 30
- XML: 5
- Total: 144
- Supplement: 14
- BibTeX: 1
- EndNote: 3
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1