Sense, Predict, Act
 

ALERTS (Automated Land change Evaluation, Reporting, and Tracking System)

ALERTS Summary

ALERTS(beta), the Automated Land change Evaluation, Reporting and Tracking System, beta edition, is a web-based prototype application for near real-time global land use and land cover change (hereafter referred to as “land change”) detection.

ALERTS(beta) can provide timely (with as little as 6-8 week latency), global coverage of deforestation or other land change events and offers users a number of useful tools for identifying, characterizing and responding to disturbances. Because it uses existing satellite data products and machine-automated change detection algorithms, ALERTS(beta) is already providing global coverage at a 1-kilometer resolution and can be readily downscaled to provide national coverage at 250m.

ALERTS(beta) is a joint effort between the Planetary Skin Institute, NASA, the University of Minnesota, and Brazil’s Instituto Nacional de Pesquisas Espaciaes (INPE) and is currently being evaluated by several national governments as a complement to their own national land use monitoring systems.

Technical Overview

ALERTS(beta) is a free, global public good that addresses land use challenges facing three types of decision-makers: government officials who want an ‘instant snapshot’ of land change at different scales; national governments and conservation stakeholders who want to be alerted of recent land changes in their areas of interest; and scientific and policy researchers who want to explore land change dynamics over a long period in greater detail, preferably in a geospatial environment.

ALERTS(beta) has three technical components: a change detection system; a geospatial analysis environment; and a web 2.0 portal that enables customized alerts of land change. Each is described below:

Change detection system: ALERTS(beta) is based on geospatial data mining algorithms, developed by the University of Minnesota, that leverage MODIS time series data. These algorithms, which draw on ten years of research in signal processing and data mining technology and have been published in peer-reviewed journals, identify meaningful patterns in vegetation signals. The ALERTS(beta) suite of algorithms can identify sudden drops, gradual decreases, or gradual increases in vegetation, and are robust to missing data, poor quality observations, and image registration issues. Moreover, they can provide both a reconstructed historical spatial record of deforestation and near real-time change detection services.

Geospatial visualization and analysis environment: ALERTS(beta) includes a web-based geospatial environment with a number of important features, including:
• Selective visualization of disturbances at multiple scales
• Direct access to vegetation reflectance time series for each disturbance to better understand the nature of land change
• Ability to compare these time series to other environmental time series such as temperature and precipitation
• Access to selected contextual layers, including protected areas, intact forests, and terrestrial carbon density, allowing pattern identification and prioritization
• Analysis tools, such as polygon aggregation, animation of historical trends, layer transparency, and carbon calculation

Web 2.0 platform: A web-based portal also allows users (i.e. governments, NGOs) to subscribe to alerts that send an automated notice whenever land change events are detected in an area of interest. Areas of interest can be defined based on contextual layers, including political boundaries and protected area status.