seth mcginnis
Seth McGinnis in his office. Seth McGinnis worked on a number of projects during FY07 related to making weather, climate, and climate-change data available to and useable by researchers, decision-makers, and the public.
NARCCAP
NARCCAP is the North American Regional Climate Change Assessment Program, an international, multi-organizational effort to generate high-resolution climate scenarios by nesting multiple regional climate models (RCMs) within multiple global circulation models (GCMs) over the coterminous United States and most of Canada.
Along with Larry McDaniel, Seth is responsible for quality-checking the expected 40+ TB of data as it is submitted for archiving. Work this year has included the development of an automated testing suite to ensure that data is correctly formatted and uncorrupted; coordinating plans with colleagues in CISL for archiving data in NCAR's Mass Store and publishing it through the Earth System Grid; development of a Data Archiving Protocol document; website and end-user support; discussion and testing to ensure that end-users' needs will be met; and collaboration with the modeling groups and quality assurance team at Iowa State University as the first datasets begin to pass through the testing and archiving pipeline.
The NARCCAP website is www.narccap.ucar.edu.
RCPM
Seth automated and developed a website for the Bayesian analysis method developed by Claudia Tebaldi and collaborators. This statistical model uses output from multiple GCMs to generate probability distributions describing the range of likelihoods of change in monthly, seasonal, or annual temperature or precipitation for a particular area at a particular time in the future, given assumptions from one of the IPCC emissions scenarios. The resulting website, rcpm.ucar.edu ( R egional C limate-Change P robabilities from M ulti-Model Ensembles), allows users to make use of global change science by getting likelihoods of change in future climate for a region of interest.
Pregenerated analyses are available for regions around the world and subregions within the United States; users can also request an analysis for a custom region. Users of this service have included representatives of the governments of the Seychelles and the Gambia; researchers interested in climate change in the Ethiopian rift valley, Tajikistan, the Kenai Peninsula in Alaska, and coastal regions in Asia; and researchers working with water utilities in Colorado and the western United States.
Estimating Water Usage and Vulnerability
Working with Bob Harriss (HARC), Seth provided support to a research project at NASA's Goddard Space Flight Center to estimate water usage for irrigation using remote sensing data. Satellite imagery can be used to estimate the amount of plant matter grown in an agricultural field between two passes. A computer model of crop growth then provides an estimate of the amount of water needed to produce that growth, and subtracting natural precipitation gives the amount of water used in irrigation.
During FY07, Seth helped define a test study region along the Rio Grande River in Texas for acquisition of data to validate the method. He procured relevant weather and soil data for this region and made them available to collaborators Lahouari Bounoua (GSFC), Shannon Franks (GSFC), and Marc Imhoff (NASA).
Disaster Dynamics: Hurricane Landfall
The Disaster Dynamics project was initiated by Robert Harriss (HARC) in 2003. Seth worked with Eric Scharff to develop an educational computer game about the interactions between human decisions and natural hazards in a fictional Gulf Coast barrier island community.
In FY07, Seth continued to maintain and update the game's website, www.dd.ucar.edu, developed new support materials, and worked with Lynette Laffea to refine the game for improved usability.
Seth facilitated the game for various groups visiting NCAR during the year, including the Hubert M. Humphrey Fellowship recipients, as well as facilitating it for a class at the DU Graduate School of International Studies. He presented a poster the game at the Weather Forecast Integration Workshop at the Naval Postgraduate School in Monterey, CA in January of 2007, and will be presenting it to the National Science Teachers Association at their regional meeting in Denver in November.
Hurricane Evacuation ABM
Evacuation is one of the most important mitigation strategies for dealing with the hazard presented by hurricanes. There are many powerful and sophisticated models of traffic used to help deal with the problems of evacuation, but even the best of those models are driven with simplistic assumptions about who will choose to leave when in response to an oncoming storm.
In FY07, Seth McGinnis worked with Sandy Johnson (DU), Brian Bush (LANL), and Brian Muller (CU-Denver) to develop a proposal for the creation of a computer model of the decision-making processes members of a population use in deciding whether and when to evacuate. This model integrates two existing statistical models of evacuation in the framework of an agent-based model (ABM), adding a social networking component. The collaborators developed a simplified proof-of-concept model and a synergistic research plan that incorporates basic social science research in the form of focus group studies at DU and policy analysis at CU-Denver with the modeling effort at NCAR.
Funding Sources
Water vulnerability research is supported by NASA. All other work is supported by the National Science Foundation, with additional funding for the RCPM project from the NCAR Weather and Climate Impacts Assessment Science Program and AwwaRF.