Leaf spectroscopy reveals drought response variation in Fagus sylvatica saplings from across the species range
Data files
Nov 21, 2024 version files 11.22 MB
-
1_LeafSpectralData.xlsx
10.20 MB
-
2_StomatalConductance.xlsx
24.49 KB
-
3_Length.xlsx
17.68 KB
-
4_SoilMoisture.xlsx
964.05 KB
-
Metadata_trees.xlsx
9.87 KB
-
README.md
1.53 KB
Abstract
The common European beech (F. sylvatica), sensitive to prolonged drought, is expected to shift its distribution with climate change. To persist in novel environments, young trees rely on the capacity to express diverse response phenotypes. Several methods exist to study drought effects on trees and their diverse adaptive mechanisms, but these are usually destructive and challenging for the large sample numbers needed to investigate biological variation.
We conducted a common garden experiment outdoors, but under controlled watering conditions, with 180 potted two-year-old saplings from 16 beech provenances across the species’ range, representing three distinct genetic clusters. Drought stress was simulated by interrupting irrigation and stomatal conductance and soil moisture were used to assess drought severity. We measured leaf reflectance of visible to short-wave infrared electromagnetic radiation to determine drought-induced changes in biochemical and structural traits derived from spectral indices and a model of leaf optical properties.
We quantified changes in pigmentation, water balance, nitrogen, lignin, epicuticular wax, and leaf mass per area in drought-treated saplings, revealing differences in likely adaptive responses to drought. Fagus sylvatica saplings from the Iberian Peninsula showed signatures of greater drought resistance, i.e., the least drought-induced change in spectrally derived traits related to leaf pigments and leaf water content.
We demonstrate that high-resolution leaf spectroscopy is an effective and non-destructive tool to assess individual drought responses that can characterize functional intraspecific variation among young beech trees. Next, this approach should be scaled up to canopy-level or airborne spectroscopy to support drought response assessments of forests.
https://doi.org/10.5061/dryad.cc2fqz6fm
In this experiment, we collected four types of data:
1) Leaf spectral data for n=180 plants (reflectances). Converted from raw .asd files to .csv using the R package spectrolab and added metadata.
2) Plant stomatal conductance for n=141 plants (in [mmol/m2s]). Provided as .xlsx file including metadata.
3) Plant height for n=180 plants (in [cm]). Provided as .xlsx file including metadata.
3) Soil moisture data (3936 time points for n=20 plants in uncalibrated Time Domain Transmission (TDT) units). Provided as .xlsx file including metadata.
Description of the data and file structure
Each data file (.xlsx) contains the necessary metadata describing the data collected, units, and how sampling and collection occurred (one single file with a tab for data and another for metadata per dataset).
The variable “site” (values BES – SLK) appearing in the data files is described in the file metadata_trees.xlsx. This file also describes the location of the sites and the date of seed collection.
The variable “collection” (values “1 - WNW”) refers to either the mother tree (integers) or exposition (E: East, N: North, W: West, S: South).
Code/Software
To process the data, we used R (v4.3.1, R core team 2023). Spectral data were processed with the spectrolab package (v0.0.18, Meireles et al 2020).
Common garden experiment
We conducted our experiment with 180 2-year-old saplings from 16 provenances across the natural distribution range in June 2023. The saplings were transported from the Hauenstein nursery to the experiment site at the University of Zurich, Zurich, Switzerland (47° 23’ 44” N, 8° 33’ 05” E; elevation 509 m.a.s.l.). We implemented a 2 x 5 randomized block arrangement, in which each block formed a 3 x 6 grid. Each block contained 18 saplings, which were placed in the slots of a steel grid, which were spaced approximately 1 m apart to minimize interactions between saplings of different blocks. We maximized the diversity of provenances within the blocks, so each provenance was represented in every block whenever possible. For each provenance, half of the saplings were assigned to the drought treatment and half to the control. The placement of the saplings within the blocks was randomized. In April 2023, we measured the length of the primary shoot using a flexible tape measure from the root collar along the shoot to the highest point in its natural position. We then used these plant height measurements to select individuals to have a similar distribution of heights among the blocks.
Drought treatment
We selected a duration of 14 days as a target to induce realistic drought stress while avoiding tree mortality in the experiment. During the simulated drought period, both groups were subjected to a controlled exclusion from rain using cone-shaped rain covers which enclosed the entire pot, but the control group was watered regularly according to demand and at least every other day. Covers were made of waterproof and UV-resistant PVC black pond foil (Heissner GmbH, Lauterbach, Hessen, Germany) and fixed on the sapling stem with Parafilm (Pechiney Plastic Packaging Inc., Chicago, IL, United States) beneath the first stem node. Three thin wooden sticks were used to elevate the cover to leave a gap between the foil and the pot to enable air circulation. In addition, we inverted the pot trays in the treatment group to avoid the retention of water in the soil. Note that we prioritized sapling survival by ad hoc irrigation of the drought-treated saplings: when severe wilting was observed, we watered wilting saplings with 600 mL water.
To monitor soil moisture and air temperature, we installed 20 TMS-4 probes (TOMST, Prag, Czech Republic, Wild et al., 2019) in pots from block A to F. The probes covered as many provenances as possible while capturing the full range of plant heights within a single block. Soil moisture and soil temperature were sampled in an interval of 15 minutes during the entire experiment.
Leaf spectroscopy
Leaf spectroscopy measurements were taken during the second half of the drought (days 9 to 14, corresponding to June 19th to 24th, 2023). We constrained the measurement time to ±3 hours from solar noon. Therefore, within one day, we measured two blocks (one control, one treatment block) totalling 36 individuals and 72 leaves (2/individual), except for one day when 72 individuals corresponding to 144 leaves were measured. We used the ASD FieldSpec 4 Standard-Res spectrometer (ASD Inc., Boulder, USA), which measures electromagnetic radiation across a total of 2151 bands covering the spectral range from 350 nm to 2500 nm. We used a plant probe with a leaf clip (model A122317, serial No 455, ASD Inc., Boulder, USA) with a calibrated low-intensity halogen light source. The leaf clip allows the non-destructive handling of the leaves and the isolation of external illumination. We used the spectral acquisition software RS3 (ASD Inc., Boulder, CO, USA). For each sapling, we measured two random sun-exposed top-of-canopy leaves. We placed the leaf clip on the leaf adaxial surface, filling the measurement window of the plant probe and avoiding the midrib. Spectral measurements consisted of four successive measurements for each leaf: the white reference (Rw), the white reference with target leaf (Tw), the black reference (Rb), and the black background reference with target leaf (Tb). We gathered 5 readings per measurement, totalling 20 measurements per leaf. We recalibrated (optimized) the sensor gain with the RS3 software against the white reference background of the leaf clip after measuring 10 saplings to correct for differences in temperature sensitivity of the three detectors resulting in deviations of the white reference reflectance value from one. The leaf clip and plant probe were regularly checked for dirt or other contamination and cleaned, or white and black reference stickers were replaced as needed.
Stomatal conductance
During the drought, we measured stomatal conductance (gs) of a subset of the saplings (n=141) with an SC-1 leaf porometer (METER Group, Pullman, WA United States) according to the manufacturer’s protocol. Along with the spectral measurements, we constrained the measurement time to ±3 hours from solar noon due to high variation during the day. Given this constraint and the significant time required for each measurement, it was not feasible to measure all the beech saplings in a single day. Therefore, we measured two to three blocks each day. To standardize the strong and immediate effects of changing conditions on stomatal conductance during the 6-hour measurement window, we always measured a sapling from the control directly after measuring one from the treatment. With this ”paired” approach, the solar radiation and air temperature were on average similar for treatment and control, which allows comparison between them. We calibrated the device according to the manufacturer’s manual on each measurement day. When measuring, we took the values as soon as they stabilized. Per sapling, we measured two representative sun-exposed leaves on the adaxial side.