Original papersSATVeg: A web-based tool for visualization of MODIS vegetation indices in South America
Introduction
In recent years, time series of satellite images have been increasingly used in a wide range of applications, especially involving the Earth's surface monitoring. Examples of using time series include the mapping of agricultural crops (Arvor et al., 2011, Picoli et al., 2018), the detection of land use and land cover (LULC) changes (Klein et al., 2012, Usman et al., 2015), the agricultural intensification studies (Kastens et al., 2017, Oliveira et al., 2014), the vegetation seasonality and phenology identification (Martínez and Gilabert, 2009, Sakamoto et al., 2005), among others. The spectral-temporal approach explores the short time of revisiting by some orbital sensors in order to obtain more often spectral information from the Earth's surface, bringing advantages over the traditional approach, based on a limited set of images.
Multi-temporal analyses of land cover monitoring are usually based on vegetation indices (VI), derived from mathematical combinations between spectral bands, which seek to enhance the response of the vegetation and decrease the soil and atmospheric influences (Jackson, 1983). VI present high correlation with the green biomass and leaf area index (LAI) and the most used are the Normalized Difference Vegetation Index (NDVI) (Rouse et al., 1974) and the Enhanced Vegetation Index (EVI) (Huete et al., 1994). When organized and observed chronologically, these indices can be used to generate long term curves, representing the green biomass variations over time, which can be related to LULC patterns and their dynamics, such as deforestations, burnings, floods, productive system changes, among others.
One of the most important sensors used in multi-temporal studies is the Moderate Resolution Imaging Spectroradiometer (MODIS), the main instrument aboard the Terra and Aqua orbital platforms of the Earth Observing System (EOS) program, led by the National Aeronautics and Space Administration (NASA). MODIS time series comprises 20 years of good radiometric and spatial quality data and is available by US Government repositories, like the Land Processes Distributed Active Archive Center (LP DAAC).
In general, time series analysis requires handling a large volume of data derived from the satellite images and involves robust computer processing. Programming languages and other computational tools for batch downloading and automatic processing are typically employed to deal with this issue. LP DAAC Data Pool repository allows batch downloading of MODIS products using specific software or programming scripts. Some applications and programming codes developed for automatic processing of MODIS products are also available and allow automatizing basic image processing steps, like format conversions, reprojection, spatial resampling and layer stacking.
However, in the last few years a sort of websites has been developed to provide the visualization and analysis of geospatial data, without needing to download, process or handle a large amount of data, such as satellites images. New geospatial database models, map servers and data visualization libraries have been developed and used to build the back-end and front-end infrastructures of websites, resulting in an increasing number of web-tools specially designed to provide fast and easy access to geoinformation. Some web-tools were available to provide visualization and analysis of MODIS VI times series (Eberle et al., 2013, Freitas et al., 2011), although issues related to response time, database updating or even the lack of continuity of these services have not been completely resolved. On the other hand, on-demand services, such as AppEEARS – Application for Extracting and Exploring Analysis Ready Samples (https://lpdaac.usgs.gov/tools/appeears/) – have also been available to provide users with a variety of remote sensing data products, including MODIS VI time series, by submitting requests to subset data spatially and temporally, which are later available for download.
Considering the increasing interest in this theme, the demands for more efficient and faster computational tools and solutions for data storage, organization, processing, publication and interoperability have also grown. The lack of available web-based systems for fast data querying on historical times series of MODIS NDVI and EVI images, including a set of specific features for data visualization, stimulated the seek for new alternatives. Thus, this article aims to report the development of the Temporal Vegetation Analysis System (SATVeg), a free online tool designed to provide instantaneous access to temporal profiles of MODIS VI in South America through a simple and friendly user interface.
Section snippets
Remote sensing data
The MODIS images used in the development of SATVeg were obtained from the LP DAAC Data Pool repository (http://lpdaac.usgs.gov). We acquired the complete time series of MOD13Q1 and MYD13Q1 products (version 6), in HDF (Hierarchical Data Format) format and sinusoidal projection, covering the South America territory. These products include 16-days image composites with spatial resolution of 250 m and geometric, radiometric and atmospheric correction levels.
MODIS composites are built from the best
Geospatial database development
Building the SATVeg database involved choosing the optimal quadrant size for storing the time series in partitioned tables. Table 1 presents the results of the performance analysis of three different database setups, taking into account the importation of 700 pairs of partitioned NDVI images in quadrants of 100x100, 150x150 and 200x200 pixels.
The smaller the quadrants size the greater the amount of partitions in the database table and, consequently, the longer the image import time. According
Conclusions
We presented the Temporal Vegetation Analysis System (SATVeg), a web-based system for fast access and visualization of time series of vegetation indices in South America. The back-end architecture was implemented through an open source relational geospatial database to store the complete time series of MODIS VI data. The strategy of using thousands of partition tables in a geospatial database to store the VI time series data ensured quick queries and efficient updating. The front-end was
CRediT authorship contribution statement
Júlio César Dalla Mora Esquerdo: Conceptualization, Software, Investigation, Validation, Writing - original draft, Writing - review & editing, Visualization, Supervision, Funding acquisition. João Francisco Gonçalves Antunes: Conceptualization, Investigation, Validation, Writing - review & editing, Supervision, Funding acquisition. Alexandre Camargo Coutinho: Conceptualization, Investigation, Validation, Resources, Writing - review & editing, Supervision, Funding acquisition. Eduardo Antonio
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
SATVeg was partially funded by Agroicone company through a technical and financial cooperation agreement between the Brazilian Agricultural Research Corporation (Embrapa) and the Arthur Bernardes Foundation (Funarbe). The authors thank NASA’s Land Processes Distributed Active Archive Center (LP DAAC) for providing the MODIS images used by SATVeg.
References (26)
- et al.
A curve fitting procedure to derive inter-annual phenologies from time series of noisy satellite NDVI data
Remote Sens. Environ.
(2007) - et al.
A simple method for reconstructing a high-quality NDVI time-series data set based on the Savitzky-Golay filter
Remote Sens. Environ.
(2004) - et al.
A simplified data assimilation method for reconstructing time-series MODIS NDVI data
Adv. Sp. Res.
(2009) - et al.
Development of vegetation and soil indices for MODIS-EOS
Remote Sens. Environ.
(1994) - et al.
TIMESAT—a program for analyzing time-series of satellite sensor data
Comput. Geosci.
(2004) - et al.
Regional land cover mapping and change detection in Central Asia using MODIS time-series
Appl. Geogr.
(2012) - et al.
Vegetation dynamics from NDVI time series analysis using the wavelet transform
Remote Sens. Environ.
(2009) - et al.
Regional yield estimation for winter wheat with MODIS-NDVI data in Shandong, China
Int. J. Appl. Earth Obs. Geoinf.
(2008) - et al.
A crop phenology detection method using time-series MODIS data
Remote Sens. Environ.
(2005) - et al.
High spatial resolution land use and land cover mapping of the Brazilian Legal Amazon in 2008 using Landsat-5/TM and MODIS data
Acta Amaz.
(2016)