The Tier 2 interface is being developed for users with basic or advanced familiarity with computer coding and ecological, earth science, and hydrological modeling. These users can operate at the coding prompt level or through basic interactive interfaces developed using the shiny web application network and the leaflet package for R (Cheng and Xie, 2017; Cheng et al., 2017). This tier will provide more flexibility in the processing of the information and a higher level of customization in the analysis.
Up to now, we have crafted two fully documented prototype versions of R packages: ecochange and rasterMapR.
Former package ecochange produces maps of forest ecosystem extent for any given time period between 2000 and 2019 in pre-defined administrative units or for user defined areas of interest in any location in the world based on Global Forest Change data. The package also produces statistics of forest extent and fragmentation and, combined with other data sets such as global tree-cover, or species distribution ranges, delivers statistics and reports on changes in essential biodiversity variables related to ecosystem structure and species populations. In-package routines include the processing of polygon geometries and the downloading, cropping and merging of Earth Observation and other spatial products and changes in the distribution range of different species. ecochange V1.3 has been freely available in the R-CRAN software repository since October, 2020, and has been widely adopted by the scientific community, exhibiting over 2,300 downloads up to date. The functionality of the ecochange package has already been incorporated into the backend system.
Second package rasterMapR has functionalities for radiometric normalization of earth observations from different sensors or acquisition dates. The functionality of the packages allows users to decompress data, intersect adjacent layers, extract Pseudo Invariant Features (PIF), and use them to express pixel values of a target image as a function of a reference one. This package facilitates the production and temporal extrapolation of temporal maps such as land cover and ecosystem extent helps to improve the quality of the inputs required by the DSS in particular to those related to mapping changes in ecosystem extent. We are expanding the functionality of the rasterMapR package to produce an automated workflow for image preprocessing, thematic map production and validation and expect to publish a stable version of this package in CRAN repository in the coming months.