Consistent spectral reflectance signatures of photosystem II thermal tolerance in contrasting foundation tree species
Data files
Mar 17, 2026 version files 103.77 MB
-
Metrosideros_polymorpha_data.xlsx
37.35 MB
-
Populus_fremontii_data.xlsx
63.45 MB
-
README.md
3.77 KB
-
run_tcrit_spectroscopy_analysis.ipynb
2.97 MB
Abstract
Photosystem II (PSII) is among the most thermally sensitive components of photosynthesis, and emerging evidence suggests that plants in diverse biomes face increasing risk of PSII damage under future climate change. However, uncertainties in the distribution and drivers of PSII thermal tolerance (Tcrit) limit our ability to predict thermal risk in plant communities across spatial scales. Here, we evaluate whether intraspecific variation in Tcrit corresponds to leaf reflectance spectra (400-2500nm) to identify mechanisms associated with Tcrit in field conditions and assess the potential of its remote estimation using remote sensing platforms. We measured Tcrit using temperature response curves of minimal fluorescence (Fo) along with corresponding leaf reflectance spectra in two foundation tree species: Populus fremontii (US Southwest) and Metrosideros polymorpha (Hawai‘i). P. fremontii was sampled under both moderate (<40ºC) and extreme (>45ºC) heat. Consistent spectral signatures of Tcrit emerged across species and sampling conditions, with the strongest signatures in P. fremontii under extreme heat. These signatures allowed Tcrit estimation (R²=0.24-0.30; RMSE<1.0ºC) and classification of high- versus low-Tcrit (71-77% accuracy) in P. fremontii. Across both species, Tcrit tended to increase with spectral indices reflecting higher chlorophyll content and lower carotenoids, nonphotochemical quenching, and leaf water content. These findings suggest that variation in PSII thermal tolerance is linked to fundamental biochemical properties of leaves, which are reflected in their optical traits. As climate extremes intensify, spectral screening and scaling of Tcrit via remote sensing may support improved conservation, management, and thermal risk assessment in vulnerable ecosystems.
Dataset DOI: 10.5061/dryad.h9w0vt4w3
Description of the data and file structure
These data were collected to determine whether leaf thermal tolerance (Tcrit) can be detected with leaf reflectance spectra in two contrasting tree species, and if so, why. Numerous analyses were performed (see methods) to determine predictive accuracy of Tcrit using only leaf spectra as well as to determine the chemical, structural, and physiological basis for any relationship.
Files and variables
File: Metrosideros_polymorpha_data.xlsx
Description: Combined leaf spectra and leaf thermal tolerance data from Metrosideros polymorpha, collected at a M. polymorpha common garden near Volcano, HI. Sheets include 'Spectra all scans' and 'Physiology data'
Variables
- LeafID -- an ID that is specific to each leaf, in the format: Dx (Day x); Px (Population x); Tx (tree x); Lx (leaf x)
- Visible albedo, NIR SWIR ALBEDO, Total Shortwave Albedo (unitless): solar-weighted reflectance for different regions, not used in this analysis.
- Site (Common garden or various in situ locations). Only common garden data was used for this analysis
- ScanNum: Unique scan number. Usually 3 per leaf.
- FvFm (ratio; unitless): Max quantum yield of PSII, measured in Hawaii
- Tmax : Temperature at which Fo peaks
- T50 (ºC): Temp at which Fo reaches 50% of difference between max and starting Fo
- Tcrit (ºC): breakpoint in Fo
- Tcrit_se (ºC): standard error in Tcrit
- SourceElevation (feet): elevation at which seeds were collected for a given population
- sourceSoil: type of soil (old or new) of that population
- ohiaGlabPub: Leaf type of population (glabrous or pubescent)
- Month: month of collection (all February for Ohia)
File: Populus_fremontii_data.xlsx
Description: Combined leaf spectra and leaf thermal tolerance data from Populus fremontii, collected at a P. fremontii common garden near Yuma, AZ. Sheets include 'Spectra all scans' and 'Physiology data'.
Variables
- Same as M. polymorpha data sheet, except without the sourceSoil or ohiaGlabPub, and SourceElevation is in meters instead of feet. Additional variables not used in analysis are from the collection presented in Moran et al. 2003 (Moran, Madeline E., et al. "Limits of thermal and hydrological tolerance in a foundation tree species (Populus fremontii) in the desert southwestern United States." New Phytologist 240.6 (2023): 2298-2311.)
File: run_tcrit_spectroscopy_analysis.ipynb
Description: This files is a Jupyter Notebook script for reading in and processing the data, as well as making the figures used in the manuscript. Each block generally makes one figure, and the purpose of each block is commented at the top of the block.
Code/software
Code used to process data, do analysis, and create figures is found in run_tcrit_spectroscopy_analysis.ipynb. This file is a Jupyter Notebook script for reading in and processing the data, as well as making the figures used in the manuscript. Each block generally makes one figure, and the purpose of each block is commented at the top of the block. Python packages used can be found at the top of the script, or at the tops of the blocks.
Access information
Other publicly accessible locations of the data:
- The P. fremontii physiology data used in this paper can be found in the supplemental information here:
- Moran, Madeline E., et al. "Limits of thermal and hydrological tolerance in a foundation tree species (Populus fremontii) in the desert southwestern United States." New Phytologist 240.6 (2023): 2298-2311.
We used two established common gardens – one for P. fremontii near Yuma, Arizona (Cooper et al., 2018) and one for M. polymorpha near Volcano, Hawaii (Cordell et al., 1998) – to assess relationships between Tcrit and leaf reflectance (400-2500 nm). We sampled M. polymorpha in a single field campaign (February 2023), but sampled P. fremontii twice (May and August 2021) to compare spectral-Tcrit relationships in the same trees under moderate (<40ºC) and extreme (>45ºC) heat (Moran et al., 2023). For each dataset, we analyzed spectral trends and indices, compared how much variation in thermal tolerance each spectral region captured, and tested the ability of reflectance spectra to predict and classify Tcrit.
Full methods can be found in the associated manuscript.