1. Global Earth Science Community
  2. Community Data System Programs
  3. REASoN Projects
  4. TOPS: Agent-Based Interface to Terrestrial Ecological Forecasting

TOPS: Agent-Based Interface to Terrestrial Ecological Forecasting

Ramakrishna Nemani, PI, NASA's Ames Research Center
Petr Votava, co-PI, California State University Monterey Bay (CSUMB)
Forrest Melton, co-PI, CSUMB
Oren Etzion, co-PI, University of Washington
Andrew Michaelis, co-PI, CSUMB

The overall goal of NASA's Terrestrial Observation and Prediction System (TOPS) Agent-Based Interface to Terrestrial Ecological Forecasting project is to develop a flexible architecture for ecological forecasting that combines multiple distributed data sources and models to provide near-real-time answers to questions about the state of the Earth system. By providing technologies for automation not only of ecological forecasting, but also of the post-processing on model outputs needed to visualize the results, TOPS will facilitate near real-time monitoring and forecasting of ecosystem conditions within U.S. and international protected areas. This capability will be essential to fully achieve the promise of a Sensor Web, in which huge volumes of real-time data are used to support activities ranging from basic science to monitoring and tracking severe weather and critical biospheric events.

Research conducted under this NASA Research, Education and Applications Solutions Network (REASoN) project is focused on the automated production and delivery of a suite of ecosystem products for use in the operational management of national parks and other protected areas. Initial product development will focus on Yosemite National Park in California, with the operational system being implemented throughout the western United States. Products currently under development use ecosystem models to integrate measurements from multiple satellite instruments, meteorological observations from multiple ground-based networks, and ancillary data sources such as soil databases and digital elevation models. The results from these models will provide data related to ecosystem function and will give managers advance warning of potential threats such as anomalies in vegetation growth and ecosystem productivity, invasive species, changes in water balance, and land cover changes both within the park and in areas adjacent to park boundaries.

These ecosystem nowcasts and forecasts will be produced automatically using TOPS and the Ecocast < architecture, which includes the IMAGEbot planner and an API for integration of new models. TOPS will be used to retrieve, pre-process, and verify model inputs, and manage the distribution of forecasts of ecosystem conditions to U.S. National Park Service scientists and staff.

Last Updated: Oct 22, 2018 at 1:38 PM EDT