Research Products  /  Open dataset

Grassland Imaging Spectroscopy and Drought Dataset

Airborne hyperspectral images, plant trait measurements and topographic data for a mixed grassland at Koffler Scientific Reserve, Ontario, in a drought year (2016) and a normal year (2017).

301Spectral bands, 400–1000 nm
~0.2 mSpatial resolution
244Leaf samples, 2016–2017
5Grassland species mapped

Overview

This dataset supports the first study to use high-resolution airborne imaging spectroscopy to characterize the variation of chlorophylls (Chl), carotenoids (Car), the Chl/Car ratio, water content (WC) and leaf area index (LAI) across topographic gradients and under drought stress, at the species level, in a heterogeneous grassland.

The study site is a grassland at Koffler Scientific Reserve in Ontario, Canada, dominated by two invasive species, smooth brome (Bromus inermis) and orchard grass (Dactylis glomerata), and three native species, goldenrod (Solidago canadensis), common milkweed (Asclepias syriaca) and wild grape (Vitis riparia). Trait values decreased from lower to upper slope positions, and drought-induced trait responses differed within and between species, particularly between the drought-tolerant invasive smooth brome and the native species.

This dataset can support grassland conservation and the management of invasive species. The images, trait measurements and maps can be used to monitor plant condition and drought stress across varied terrain and to predict where invasive grasses such as smooth brome are likely to establish and spread, as well as to develop methods for mapping plant traits and species with airborne and satellite imaging spectroscopy, such as the EnMAP and PRISMA missions and the future Surface Biology and Geology mission.

Random Forest · 2017 image

Species map

Object-based classification of the five species, trained and tested on 350 reference samples: 89.6% overall accuracy, kappa 0.870.

Partial least squares regression

Trait maps

Chl, Car, Chl/Car, WC and LAI mapped from the airborne images for both years; leaf-level models were built from leaf reflectance measured with a FieldSpec 3 spectroradiometer.

2016 versus 2017

Drought response

Trait differences between the drought and normal years across six topographic categories, from valley to ridge. WC and LAI changed most; Chl/Car was the most stable.

Study details are from the dataset README and the preprint of the article (bioRxiv, 2024).

Fieldwork

Photos from related RSSEM Lab field work; they are not part of the Dryad dataset.

What’s included

  • Hyperspectral images20160823_HSI.tif and 20170820_HSI_INT_N.tif (GeoTIFF): 301 bands, 400–1000 nm at 2 nm; 2017 reflectance scaled by 10,000 and stored as integers
  • Species classification dataTraining (80%) and test (20%) shapefiles, one per species, and a soil and shadow mask for 2017 (20170820_soil_shadow_mask.tif)
  • Trait and spectral tablesLeaf-level (350–2500 nm) and image-level (400–1000 nm) spectra with Chl, Chl a/b, Car, Chl/Car, WC and SLA or LAI, for 2016 and 2017 (CSV)
  • Topographic position indextpi_2m_rescale_clip.tif: TPI raster of the study area
  • Ground sampling sitesGroundSites_2016.shp and GroundSites_2017.shp: locations of the field sampling sites
  • Analysis tablesSoil moisture and TPI at 42 field sites; canopy traits against TPI; traits and 2016–2017 trait differences by six topographic categories (CSV)
  • CodeSpecies_classification_RandomForest.ipynb (Python); PLSR_leaf_model.R, PLSR_image_model.R and Statistical_analysis.R
  • Files on DryadData_and_Codes.zip (6.67 GB) and README.md (12.4 kB); published 3 January 2025
  • LicenceCC0 1.0 Universal (public domain dedication)
  • FundingNatural Sciences and Engineering Research Council of Canada, Discovery Grant RGPIN-386183
Access the dataset

Cite

Dao, P. D., He, Y., Lu, B., & Axiotis, A. (2025). Imaging spectroscopy reveals topographic variability effects on grassland functional traits and drought responses. Ecology, 106(3), e70006. https://doi.org/10.1002/ecy.70006

Dao, P. D., He, Y., Lu, B., & Axiotis, A. (2025). Imaging spectroscopy reveals topographic variability effects on grassland functional traits and drought responses [Data set]. Dryad. https://doi.org/10.5061/dryad.2jm63xt0d

Please cite the article and the dataset when using these data.

Contact

Questions about the data, the study site or collaboration on imaging spectroscopy of vegetation are welcome. Please get in touch with the laboratory.

For other enquiries, see the lab contact page.

Prof. Yuhong He

Principal Investigator, RSSEM Lab
Department of Geography, Geomatics and Environment, University of Toronto Mississauga

Office
DV3271, William G. Davis Building
Address
3359 Mississauga Road, Mississauga, ON L5L 1C6
Email about this dataset