Site ID:
US-PR0; XX-PTr
Site Name:
Island of Puerto Rico; Pantropical site
Variables:
Leaf area index; Precipitation (rainfall); Remote sensing; Soil chemical composition; Soil types (or soil order); Wind speed/direction
Date Range:
Feb. 29, 1988
-
Sept. 20, 2017
Description:
Statement of purpose: Cyclones alter the function and composition of tropical forests, making effects of intensifying cyclones on carbon-rich forests a critical topic of study. Here, we quantified cyclone-induced damage and recovery of 21 cyclone disturbances affecting 23 pantropical forest sites between 1988-2017 utilizing leaf area index (LAI), enhanced vegetation index (EVI), normalized difference vegetation index (NDVI), and transformed NDVI (kNDVI) values from Google Earth Engine. Field observations collected in a meta-analysis (Bomfim et al., 2022, in review) were used to ground-truth and test effects of soil resource availability and disturbance factors on damage and recovery. This meta-analysis also served as the basis to begin vegetation index extraction, utilizing unique site and date combinations, from tropical forests effect by cyclone disturbances. We began collecting NDVI (5km resolution) from the NOAA Climate Data Record (CDR) of AVHRR Normalized Difference Vegetation Index (NDVI), Version 5 data product (Vermote, 2019) for all case studies included, 42. Next, we began extracting Landsat data from Landsat 4, 5, and 8, courtesy of the U.S. Geological Survey, in search of higher resolution data. We selected a 3 by 3 Landsat pixel area, leading to a 90m resolution data extraction. The specific imagery used includes Landsat 4 USGS Landsat 4 TM Collection 1 Tier 1 TOA (top of atmosphere) Reflectance, Landsat 5 USGS Landsat 5 TM (thematic mapper) Collection 1 Tier 1 TOA Reflectance, and Landsat 8 USGS Landsat 8 Collection 1 Tier 1 TOA Reflectance. Within Google Earth Engine, we selected the date and location (latitude and longitude), calculated NDVI, kNDVI, and EVI utilizing Landsat bands (see metadata_NGEE-tropics_cyclones), and extracted post- and pre-cyclone values for each case study to calculate cyclone-induced change in the vegetative index. Due to limited spatial resolution of Landsat remote sensing data, MODIS products were investigated next. First, the MOD13Q1.006 Terra Vegetation Indices 16-Day Global 250m product was used to extract 250m EVI and NDVI (Didan, 2015) and then the MCD15A3H.006 MODIS Leaf Area Index/FPAR 4-Day Global 500m product product was used to extract LAI 500m (Myneni et al., 2015). Pre- and post-cyclone values, change in the vegetative index, and standard deviation for all values are included in the main csv (see case_study_data.csv) for all vegetative indices collected, including LAI 500m, EVI 250m, NDVI 250m, NDVI 90m, kNDVI 90m, EVI 90m, and NDVI 5km. Lastly, recovery values were calculated utilizing a standardization method (see metadata_NGEE-tropics_cyclones) and recovery values for MODIS (see MODIS_recovery.csv) and Landsat (Landsat_recovery.csv) data are included.
QA/QC:
None
Methods Description:
Detailed methods are described in the Metadata file below tables 1-3 (see metadata_NGEE-tropics_cyclones).
Find general methods below:
Site Selection
This study used 42 case studies extracted from a meta-analysis focused on cyclone-induced litterfall responses within tropical latitudes. The meta-analysis and its components allowed for comparisons between litterfall (total and leaf), remote sensing derived LAI, and site-specific soil phosphorus concentration. The litterfall and cyclone- and site-specific data from the meta-analysis is found in the case_study_data.csv.
Remote Sensing
Vegetative Indices Acquisition
After data products and imagery were found and obtained on Google Earth Engine, vegetative indices were either extracted directly from data products with bands for the associated vegetative index or extracted through a calculation-based method utilizing needed bands.
To apply the given case-studies to the remote sensing data to obtain the vegetative index pre- and post-cyclone data, the longitude and latitude of each site and date of the cyclone disturbance were applied in the Google Earth Engine code (coordinates and dates of extraction found in case_study_data.csv). Full syntax for the Google Earth Engine code on Github link.
Change in Vegetative Index calculation
After extracting the vegetative indices for all case studies, we calculated two ΔLAI and ΔEVI values differing in their pre-cyclone LAI calculations. This was done for all vegetative indices collected.
For recovery data, the recovering vegetative index was subtracted by the pre values and normalized to (or divided by) this value as well (data in Landsat_recovery.csv and MODIS_recovery.csv). Recovery data was found from 1-25 months post-cyclone and calculations were completed on a monthly basis, with multiple observations for one month being averaged.
Access Level:
Public
Originating Institution(s):
Climate and Ecosystem Sciences Division, Lawrence Berkeley National Laboratory; Energy & Resources Group, University of California, Berkeley; Environmental Engineering Sciences, University of Florida; Department of Geography, University of California, Berkeley
Sponsor Organization(s):
DOE SULI program at LBNL, NGEE-Tropics at LBNL, and NGEE-Tropics at UC, Berkeley
Contact:
Bloom, Dellena - Environmental Engineering Sciences, University of Florida (develyn.bloom@ufl.edu)
Version:
1.0
Dataset Citation:
Bloom D; Bomfim B; Feng Y; Kueppers L (2022): LAI, EVI, NDVI, and kNDVI in 23 pantropical forests affected by 21 cyclones. 1.0. NGEE Tropics Data Collection. (dataset). http://dx.doi.org/10.15486/ngt/1847332
Acknowledgement:
This work was supported in part by the U.S. Department of Energy, Office of Science, Office of Workforce Development for Teachers and Scientists (WDTS) under the Science Undergraduate Laboratory Internship (SULI) program.
This research was supported as part of the Next Generation Ecosystem Experiments-Tropics (NGEE), funded by the U.S. Department of Energy, Office of Science, Office of Biological and Environmental Research under contract number.
Landsat 4, 5, and 8 data courtesy of the U.S. Geological Survey.
Data Link: Download Dataset
I have read and agree to the terms of the
NGEE
Tropics data policy .
1. Bloom, D. E., Bomfim, B., Feng, Y. & Kueppers, L. M. Combining field and remote sensing data to estimate forest canopy damage and recovery following tropical cyclones across tropical regions. Environmental Research: Ecology 2, 035004 (2023). DOI: 10.1088/2752-664X/acfaa3
