The Research Data Exploratory Team and members of the Software Architecture Working Group collaborated with Professor Robert Kopp and D.J. Rasmussen to ingest and provide access to an impressive collection of county-level climate change projection data that is now available in the RUcore Research Data Portal at http://dx.doi.org/doi:10.7282/T3SF2Z93.
Rasmussen, Kopp, and collaborator Malte Meinshausen generated projections of future temperature, precipitation, and humidity by combining the probability of different global average temperature outcomes over the 21st century with spatially detailed projections from several state-of-the-art global climate models. The projections also include daily to multi-year weather variability, which is needed for economic models that estimate the impacts of climate change. The resulting data set is a 1.3 TB product that is freely available for use by researchers, decision makers, and climate change communicators. The need for local climate change projections is growing as decision makers are increasingly demanding estimates of the economic costs of future climate change and the value of avoiding associated damages.
The climate projections from Rasmussen, Kopp, and Meinshausen were used in the book Economic Risks of Climate Change: an American Prospectus. This prospectus provides a climate risk assessment that estimates the economic impact of climate change on the U.S. and provides local estimates of economic risks in multiple sectors of the U.S. economy, including labor, agriculture, and energy. The complete dataset in RUcore provides open access to all data and methodology for the physical climate projections so that results can be reproduced and improved in future studies. The technical analysis in the prospectus was commissioned by the Risky Business Project. This effort was led by former New York City mayor Michael Bloomberg, former Bush administration treasury secretary Hank Paulson, and former hedge-fund manager Tom Steyer. The aim of these three business leaders was to inspire risk managers in the business community to incorporate climate change related financial risks in their decision making process.
The RUcore repository architecture provides a set of unique features that enables researchers to easily access the parts of this dataset that are most important. Each major directory, of which there are twelve, has its own Digital Object Identifier and can be individually cited. Perhaps most useful is that file and directory names are preserved and the user can walk the directory tree to select individual files and directories for download. This feature is important since downloading a complete directory in the order of 200GB will take hours. As an alert to prospective users, we provide an estimate of how long it will take to download the requested files and directories. In addition, one of the directories includes the software for processing the data, enabling users to repeat or augment the original authors’ findings. As part of the Libraries’ exploratory process, we learned a great deal about how to ingest and manage large datasets greater than 100GB. We had to revise our memory management strategies to accommodate directories with thousands of files. As part of our process, we validated the transfer of the 1.3TB dataset from the original site to the RUcore server to insure that there were no corruptions in the transfer process. All in all, we believe that this dataset and the access to it provided by Libraries will significantly contribute to the ongoing research in climate science.
