markjohnsonubc/modistools

R package - retrieving & using MODIS data from NASA's ORNL DAAC archive

MODISTools

R package - retrieving & using MODIS data from NASA LPDAAC archive

MODISTools is an R package for retrieving and using MODIS data subsets using ORNL DAAC web service (SOAP) for subsetting from Oak Ridge National Laboratory (ORNL). It provides a batch method for retrieving subsets of MODIS http://modis.gsfc.nasa.gov/ remotely sensed data and processing them to a format ready for user-friendly application in R, such as statistical modelling.

The most important function is MODISSubsets, for requesting subsets from a given MODIS product for multiple time-series. Each time-series is defined by a coordinate location (WGS-1984), a surrounding area of interest, and a start and end date. Automating this as a batch process reduces time, effort, and human error. Alternatively, MODISTransects expands upon MODISSubsets by extracting MODIS data along a transect, and its surrounding neighbourhood. Downloaded subsets are saved in ASCII files, but can be converted to ASCII grid files for use in a GIS environment. MODISSummaries computes summary statistics of downloaded subsets and organises summarised data back with the original input dataset, creating a csv file that can be easily used for modelling; this provides efficient storage of data and a transparent process from data collection, to processing, to a form that is ready for final use.

The functions were originally used for downloading vegetation indices data, but have been generalised to provide a package that performs the same functionality for any MODIS data that available through the web service. For a list of available MODIS products, see http://daacweb-dev.ornl.gov/MODIS/MODIS-menu/products.html. Other minor functions -- including a lat-long coordinate conversion tool -- are included to aid this process.

Recent stable releases of this package have been checked and built on Windows, Mac, and Linux, and last checked on R 3.1.2 on 2014-12-22. MODISTools is written by Sean Tuck and Helen Phillips. This package can be used under the terms of the GNU GPLv3 license; feel free to use and modify as you wish, but please cite our work where appropriate. To cite MODISTools in publications, please use:

Tuck, S.L., Phillips, H.R.P., Hintzen, R.E., Scharlemann, J.P.W., Purvis, A. and Hudson, L.N. (2014) MODISTools -- downloading and processing MODIS remotely sensed data in R. Ecology and Evolution, 4 (24), 4658--4668. DOI: 10.1002/ece3.1273.

Some of the changes in recent updates:

• New function, MODISGrid, that takes downloaded ASCII files and converts them into ASCII grid files, which can be used in a GIS environment.
• Optional time-series plots for diagnostics as output from MODISSummaries.
• MODISGrid now writes MODIS projection (PRJ) files for all ASCII raster grids, so their correct projection is defined. These files can now be loaded directly into a GIS environment.
• MODISGrid now more flexibly deals with data files that contain multiple products.
• Citation for publication in Ecology and Evolution added.
• Documentation and 'Using MODISTools' vignette updated.

Changes coming soon:

• More product-specific time-series summaries in MODISSummaries.

Installation

The most recent stable release can be installed from the CRAN repository, by running:

install.packages("MODISTools")

This is the recommended way to download.

Alternatively, the most up-to-date in-development version of MODISTools can be installed from this github repository. To do this, install the package devtools if you haven't already done so.

install.packages("devtools")
library(devtools)

Then use install_github, with this repository name, to install MODISTools straight from github.

install_github("seantuck12/MODISTools", build_vignettes=TRUE)

Dependencies

http://www.omegahat.org/RCurl

http://www.omegahat.org/RSXML