Skip to main content
Dryad

Parcel level temporal variance of remotely sensed spectral reflectance predicts plant diversity

Cite this dataset

Rossi, Christian; McMillan, Nicholas; Hauser, Leon (2024). Parcel level temporal variance of remotely sensed spectral reflectance predicts plant diversity [Dataset]. Dryad. https://doi.org/10.5061/dryad.fxpnvx100

Abstract

Over the last two decades, considerable research has built on remote sensing of spectral diversity to assess plant diversity. The spectral variation hypothesis (SVH) proposes that spatial variation in reflectance data of an area is positively associated with plant diversity. While the SVH has exhibited validity in dense forests, it performs poorly in highly fragmented and temporally dynamic agricultural landscapes covered mainly by grasslands. Such underperformance can be attributed to the mosaic-like spatial structure of human-dominated landscapes with fields in varying phenological and management stages. Therefore, we argued for re-evaluating SVH's flawed window-based spatial analysis and underutilized temporal component. In particular, In particular, we captured the spatial and temporal variation in reflectance and assessed the relationships between spatial and temporal components of spectral diversity and plant diversity at the parcel level as a unit that relates to management patterns. Our investigation spanned three grasslands on two continents covering a wide spectrum of agricultural usage intensities. To calculate different components of spectral diversity, we used multi-temporal spaceborne Sentinel-2 data. We showed that plant diversity was negatively associated with the temporal component of spectral diversity across all sites. In contrast, the spatial component of spectral diversity was related to plant diversity in sites with larger parcels. Our findings highlighted that in agricultural landscapes, the temporal component of spectral diversity drives the spectral diversityplant diversity associations. Consequently, our results offer a novel perspective for remote sensing of plant diversity globally.

README: Spectral and species diversity on a parcel level

Description of the file structure

species_spectral_diversity.csv

includes the spectral richness per parcel and the calculated spectral diversity components including the temporal, spatial and spatio-temporal components for three different grassland sites in Switzerland, Netherlands and Oklahoma (United States).

  • ID: The ID of the parcel
  • Species_richness: The species richness per parcel corresponded to the total count of observed plant species within all the given parcel plots. For each study area, a different sampling design was used.
  • Parcel_size: The size in m2 of the parcel after negative buffering of 50 m for Oklahoma and 10 m in the other locations.
  • Study_area: The state where the field and remote sensing data were collected.
  • Lat_centroid: Latitude of the parcel centroid (EPSG:4326)
  • Lon_centroid: Longitude of the parcel centroid (EPSG:4326)
  • Sampling_year: The year of in-situ and remote sensing data collection.
  • Temporal_component: The temporal component of spectral diversity was calculated as the spectral variance of the average parcel value over a growing season from three Sentinel-2 vegetation indices.
  • Spatial_component: The spatial component of spectral diversity a parcel accounted for the spectral variance between pixels after averaging their temporal variability. As spectral feature three vegetation indices derived from Sentinel-2 data were used.
  • Spatio temporal: The spatio-temporal component of spectral diversity quantified the spectral variance among pixels over both space and time within a parcel, encompassing the variance not captured by the other two components.
  • Total spectral diversity: Is the sum of the three spectral diversity components per parcel.

The data was used to produce Figure 3 and the analysis in section 3.2 in Rossi et al. (2024).


specdiv_parcel.R

provides the R code to calculate the spectral diversity components on a parcel level based on multi-temporal Sentinel-2 images. In addition to the remote sensing raster files, the code requires a vector file representing the parcels as input.

The script was created using the R version 4.2.0.


spectral_diversity_CH.csv

includes the management information per parcel and the calculated spectral diversity components for the larger Swiss study site, based on Sentinel-2 data from the growing season 2021.

  • ID: The ID of the parcel
  • NUTZUNGSAR: Each grassland parcel was associated with one of seven distinct management types: (1) artificial meadows, (2) permanent meadows, (3) low-intensity meadows, extensively used meadows, (5) pastures, (6) extensively used pastures, and (7) protected grasslands
  • Lat_centroid: Latitude of the parcel centroid (EPSG:4326)
  • Lon_centroid: Longitude of the parcel centroid (EPSG:4326)
  • Shape_Leng: The perimeter in m of the parcel after negative buffering of 10 m.
  • Shape_Area: The size in m2 of the parcel after negative buffering of 10 m.
  • Temporal component: The temporal component of spectral diversity was calculated as the spectral variance of the average parcel value over a growing season from three Sentinel-2 vegetation indices.
  • Spatial component: The spatial component of spectral diversity a parcel accounted for the spectral variance between pixels after averaging their temporal variability. As spectral feature three vegetation indices derived from Sentinel-2 data were used.
  • Spatiotemporal component: The spatio-temporal component of spectral diversity quantified the spectral variance among pixels over both space and time within a parcel, encompassing the variance not captured by the other two components.
  • Total: Is the sum of the three spectral diversity components per parcel.

 
The data was used to produce Figure 2 and the analysis in section 3.1 of Rossi et al. (2024).


plot_fig_2_3.R

provides the R code to produce Figures 2 and 3 in Rossi et al. (2024). The script was created using the R version 4.2.0.

Code/Software

To perform the Welch's ANOVA test and post-hoc Games-Howell test in section 3.1,
we used the following R code:

library(onewaytests)
library(userfriendlyscience)

data_parcel_CH <-
   read.table(
     "spectral_diversity_CH.csv,
     sep = ";",
     header = TRUE
   )

welch.test(data_parcel_CH$`Spatial component`~NUTZUNGSAR,data=data_parcel) 
welch.test(data_parcel_CH$`Temporal component`~NUTZUNGSAR,data=data_parcel) 
welch.test(data_parcel_CH$`Spatiotemporal component`~NUTZUNGSAR,data=data_parcel) 

#Games Howell posthoc test our data  not meet the homogeneity of variances assumption
#tells you specific group difference

posthocTGH(y=data_parcel_CH$`Spatial component`,x=data_parcel$NUTZUNGSAR, method="games-howell",
                     conf.level = 0.95, digits=2,
                     formatPvalue = TRUE)

posthocTGH(y=data_parcel_CH$`Temporal component`,x=data_parcel$NUTZUNGSAR, method="games-howell",
           conf.level = 0.95, digits=2,
           formatPvalue = TRUE)

Methods

We calculated the spectral diversity and its spatial, temporal and spatio-temporal components from multi-temporal Sentinel-2 data in three grasslands utilized for agricultural production, two in Europe and one in the United States, using the function divcom from the stdiversity v1.1.0 package (Rossi et al., 2021). For the same parcels, we collected in situ species inventories and used species richness to express the plant diversity of each parcel. A detailetd description of data collection and processing, as well as the method used, can be found in Rossi et al. (2024).

 

Funding

European Space Agency