skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • Unveiling the transferability of PLSR models for leaf trait estimation: lessons from a comprehensive analysis with a novel global dataset
Title: Unveiling the transferability of PLSR models for leaf trait estimation: lessons from a comprehensive analysis with a novel global dataset
Leaf traits are essential for understanding many physiological and ecological processes. Partial least squares regression (PLSR) models with leaf spectroscopy are widely applied for trait estimation, but their transferability across space, time, and plant functional types (PFTs) remains unclear.We compiled a novel dataset of paired leaf traits and spectra, with 47 393 records for > 700 species and eight PFTs at 101 globally distributed locations across multiple seasons. Using this dataset, we conducted an unprecedented comprehensive analysis to assess the transferability of PLSR models in estimating leaf traits.While PLSR models demonstrate commendable performance in predicting chlorophyll content, carotenoid, leaf water, and leaf mass per area prediction within their training data space, their efficacy diminishes when extrapolating to new contexts. Specifically, extrapolating to locations, seasons, and PFTs beyond the training data leads to reducedR2(0.12–0.49, 0.15–0.42, and 0.25–0.56) and increased NRMSE (3.58–18.24%, 6.27–11.55%, and 7.0–33.12%) compared with nonspatial random cross‐validation. The results underscore the importance of incorporating greater spectral diversity in model training to boost its transferability.These findings highlight potential errors in estimating leaf traits across large spatial domains, diverse PFTs, and time due to biased validation schemes, and provide guidance for future field sampling strategies and remote sensing applications.  more » « less
Award ID(s):
2021898 1638720
PAR ID:
10518014
Author(s) / Creator(s):
; ; ; ; ; ; ; ; ; ; ;
Publisher / Repository:
New Phytologist Foundation
Date Published:
Journal Name:
New Phytologist
Volume:
243
Issue:
1
ISSN:
0028-646X
Page Range / eLocation ID:
111 to 131
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; ; ; (, New Phytologist)
    Summary Reflectance spectroscopy is a rapid method for estimating traits and discriminating species. Spectral libraries from herbarium specimens represent an untapped resource for generating broad phenomic datasets across space, time, and taxa.We conducted a proof‐of‐concept study using trait data and spectra from herbarium specimens up to 179 yr old, alongside data from recently dried and pressed leaves. We validated model accuracy and transferability for trait prediction and taxonomic discrimination.Trait models from herbarium spectra predicted leaf mass per area (LMA) withR2 = 0.94 and %RMSE = 4.86%. Models for LMA prediction were transferable between herbarium and pressed spectra, achievingR2 = 0.88, %RMSE = 8.76% for herbarium to pressed spectra, andR2 = 0.76, %RMSE = 10.5% for the reverse transfer. Discriminant models classified leaf spectra from 25 species with 74% accuracy, and classification probabilities were significantly associated with several herbarium specimen quality metrics.The results validate herbarium spectral data for trait prediction and taxonomic discrimination, and demonstrate that trait modeling can benefit from the complementary use of pressed‐leaf and herbarium‐leaf spectral datasets. These promising advancements help to justify the spectral digitization of plant biodiversity collections and support their application in broad ecological and evolutionary investigations. 
    more » « less
  2. ; ; ; ; ; ; ; ; ; ; et al (, New Phytologist)
    Summary Leaf dark respiration (Rdark), an important yet rarely quantified component of carbon cycling in forest ecosystems, is often simulated from leaf traits such as the maximum carboxylation capacity (Vcmax), leaf mass per area (LMA), nitrogen (N) and phosphorus (P) concentrations, in terrestrial biosphere models. However, the validity of these relationships across forest types remains to be thoroughly assessed.Here, we analyzedRdarkvariability and its associations withVcmaxand other leaf traits across three temperate, subtropical and tropical forests in China, evaluating the effectiveness of leaf spectroscopy as a superior monitoring alternative.We found that leaf magnesium and calcium concentrations were more significant in explaining cross‐siteRdarkthan commonly used traits like LMA, N and P concentrations, but univariate trait–Rdarkrelationships were always weak (r2 ≤ 0.15) and forest‐specific. Although multivariate relationships of leaf traits improved the model performance, leaf spectroscopy outperformed trait–Rdarkrelationships, accurately predicted cross‐siteRdark(r2 = 0.65) and pinpointed the factors contributing toRdarkvariability.Our findings reveal a few novel traits with greater cross‐site scalability regardingRdark, challenging the use of empirical trait–Rdarkrelationships in process models and emphasize the potential of leaf spectroscopy as a promising alternative for estimatingRdark, which could ultimately improve process modeling of terrestrial plant respiration. 
    more » « less
  3. ; ; ; ; ; ; ; ; ; ; et al (, New Phytologist)
    Summary Grasses are exceptionally productive, yet their hydraulic adaptation is paradoxical. Among C3grasses, a high photosynthetic rate (Aarea) may depend on higher vein density (Dv) and hydraulic conductance (Kleaf). However, the higherDvof C4grasses suggests a hydraulic surplus, given their reduced need for highKleafresulting from lower stomatal conductance (gs).Combining hydraulic and photosynthetic physiological data for diverse common garden C3and C4species with data for 332 species from the published literature, and mechanistic modeling, we validated a framework for linkages of photosynthesis with hydraulic transport, anatomy, and adaptation to aridity.C3and C4grasses had similarKleafin our common garden, but C4grasses had higherKleafthan C3species in our meta‐analysis. Variation inKleafdepended on outside‐xylem pathways. C4grasses have highKleaf : gs, which modeling shows is essential to achieve their photosynthetic advantage.Across C3grasses, higherAareawas associated with higherKleaf, and adaptation to aridity, whereas for C4species, adaptation to aridity was associated with higherKleaf : gs. These associations are consistent with adaptation for stress avoidance.Hydraulic traits are a critical element of evolutionary and ecological success in C3and C4grasses and are crucial avenues for crop design and ecological forecasting. 
    more » « less
  4. ; ; ; ; ; ; ; ; ; (, New Phytologist)
    Summary Leaf mass per area (LMA) is a key plant trait, reflecting tradeoffs between leaf photosynthetic function, longevity, and structural investment. Capturing spatial and temporal variability in LMA has been a long‐standing goal of ecological research and is an essential component for advancing Earth system models. Despite the substantial variation in LMA within and across Earth's biomes, an efficient, globally generalizable approach to predict LMA is still lacking.We explored the capacity to predict LMA from leaf spectra across much of the global LMA trait space, with values ranging from 17 to 393 g m−2. Our dataset contained leaves from a wide range of biomes from the high Arctic to the tropics, included broad‐ and needleleaf species, and upper‐ and lower‐canopy (i.e. sun and shade) growth environments.Here we demonstrate the capacity to rapidly estimate LMA using only spectral measurements across a wide range of species, leaf age and canopy position from diverse biomes. Our model captures LMA variability with high accuracy and low error (R2 = 0.89; root mean square error (RMSE) = 15.45 g m−2).Our finding highlights the fact that the leaf economics spectrum is mirrored by the leaf optical spectrum, paving the way for this technology to predict the diversity of LMA in ecosystems across global biomes. 
    more » « less
  5. ; ; ; ; ; ; ; ; ; ; et al (, New Phytologist)
    Summary Allocation of leaf phosphorus (P) among different functional fractions represents a crucial adaptive strategy for optimizing P use. However, it remains challenging to monitor the variability in leaf P fractions and, ultimately, to understand P‐use strategies across diverse plant communities.We explored relationships between five leaf P fractions (orthophosphate P, Pi; lipid P, PL; nucleic acid P, PN; metabolite P, PM; and residual P, PR) and 11 leaf economic traits of 58 woody species from three biomes in China, including temperate, subtropical and tropical forests. Then, we developed trait‐based models and spectral models for leaf P fractions and compared their predictive abilities.We found that plants exhibiting conservative strategies increased the proportions of PNand PM, but decreased the proportions of Piand PL, thus enhancing photosynthetic P‐use efficiency, especially under P limitation. Spectral models outperformed trait‐based models in predicting cross‐site leaf P fractions, regardless of concentrations (R2 = 0.50–0.88 vs 0.34–0.74) or proportions (R2 = 0.43–0.70 vs 0.06–0.45).These findings enhance our understanding of leaf P‐allocation strategies and highlight reflectance spectroscopy as a promising alternative for characterizing large‐scale leaf P fractions and plant P‐use strategies, which could ultimately improve the physiological representation of the plant P cycle in land surface models. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email