Nº 6 2015 > Celebrating ITU’s 150 Years
Big data and innovation
Innovating with Dan:
How does big data help us understand climate change?
Climate change is a real threat to our planet, with destructive consequences already being felt across the globe. Daniel Duffy, High-Performance Computing Lead at the NASA Centre for Climate Simulation (NCCS) explains how big data is essential to the task of studying climate change.
What kid hasn’t dreamed of being an astronaut? Science and space exploration have always been extremely interesting to me, so working at NASA is not only an honour and a privilege, but a dream come true. At the NASA Centre for Climate Simulation (NCCS), I work with former astronauts and some of the world’s best scientists and engineers to develop solutions to model climate change.
NCCS provides high-performance computing, storage and networking resources for large-scale NASA science projects. Many of these projects involve global simulations of the Earth’s weather and climate. These simulations create huge amounts of data; data that is literally too big for the scientists to read through. Therefore, it is increasingly important to provide new methods for analysing and visualizing the big data sets in order to better understand critical scientific questions such as climate change.
Performing research on weather forecasts will provide us with more accurate information on extreme weather events, such as hurricanes and strong convective systems that can create tornados, which have a direct impact on both the United States and the rest of the world.
Big data and climate change: How does it work?
Big data and studying climate go hand-in-hand; you really cannot study climate without large amounts of data.
The NASA Centre for Climate Simulation (NCCS) facility is home to a cluster of computers called the Discover supercomputer. Its primary objective is to provide the necessary high-performance computing and storage environment to meet the requirements of NASA science projects.
However, providing a large amount of data to a scientist means nothing if they are not able to visualize and map it in meaningful ways. One example of this is the enhanced animation produced by NASA’s Global Modelling and Assimilation Office (GMAO), which uses observations from multiple sources to drive weather forecasts.
GMAO’s GEOS‑5 Data Assimilation System (DAS) blends observational and model information to produce the most accurate and physically consistent picture of the atmosphere at any given time. Over five million observations are accumulated every six hours, mapping variables such as temperature, water, wind, surface pressure, and ozone. Assimilated observations come in eight major types, each measuring different variables from various sources.
Data mapping: Climate change and forecasting
The data generated within the NCCS contributes to a variety of key research and policy papers.
This data enables a more informed conversation on the impacts of climate change on our planet, and can help policy makers develop appropriate strategies and actions in response to climate projections. For example, the data has been used in the assessment reports sponsored by the Intergovernmental Panel on Climate Change (IPCC).
A data simulation produced in the NCCS and visualized by NASA’s Scientific Visualization Studio, presents outputs from climate models produced in the IPCC’s Fifth Assessment Report, showing how temperature and precipitation are projected to change throughout the 21st Century.
We also generate a re-analysis of the climate over the past 35 years that is used in a variety of projects outside of NASA.
Hurricane Katrina, which hit the United States Gulf Coast in 2005, highlights the importance of accurate forecasting. Though damage was immense, it could have been far worse without the forecasts to provide advanced warning and allow time for adequate preparations. Today, NCCS supercomputers host GMAO’s global circulation model, which is up to ten times the resolution of those used during Hurricane Katrina, allowing for a more accurate look inside the hurricane helping to more accurately estimate its intensity and size. This means that meteorologists have a better understanding of where a hurricane is going and what activity is happening inside, which is essential to planning successful preparations for an extreme event like Katrina.
In addition, the output of the global climate models is also used by Observing System Simulation Experiments (OSSEs) to simulate the next generation of remote sensing platforms being proposed by NASA. This provides scientists and engineers with a virtual Earth in which to study the benefits of making new remote observations of the atmosphere from space prior to even building a new sensor or satellite.
This article is an abridgement.
For full text see: http://itu150.org/story/october/