The Arctic System Reanalysis (ASR)

Description:

The Arctic System Reanalysis (ASR), which can be viewed as a blend of modeling and observations, will provide a high resolution description in space (10 km) and time (3 h) of the atmosphere-sea ice-land surface system of the Arctic. The ASR, which has been approved as an IPY (International Polar Year) full proposal under the international CARE/ASR (Climate of the Arctic and its Role for Europe/ASR) activity, and endorsed by the international World Climate Research Programme Climate and Cryosphere effort, will ingest historical data streams along with measurements of the physical components of the Arctic Observing Network developed as part of the IPY. Gridded fields from the ASR, such as temperature, radiation and winds, will also serve as drivers for coupled ice-ocean, land surface and other models, and will offer a focal point for coordinated model inter-comparison efforts. The ASR will permit reconstructions of the Arctic system's state, thereby serving as a state-of-the-art synthesis tool for assessing Arctic climate variability and monitoring Arctic change. As such, it responds directly to needs identified in the Implementation Plan for the Study of Environmental Arctic Change (SEARCH). The ASR will also shape the legacy observing network of the IPY by providing a vehicle for observing system sensitivity studies recommended in the National Research Council's (NRC) report on the design of an integrated Arctic Observing Network. The first generation ASR as will span the years 2000-2010. It will then be relatively straightforward to extend the reanalysis to earlier years and provide systematic updates.

The ASR will provide a resource for the detection and diagnosis of change in the coupled Arctic climate system. The widespread applications of existing global reanalyses (e.g., ERA-40, NCEP/NCAR, JRA-25, NASA DAS) demonstrate the high impact that can be expected on Arctic research. The system-oriented approach required for the ASR will lead to interdisciplinary collaboration, involving the Arctic terrestrial, sea ice, ocean and atmospheric communities. The reanalysis will capitalize on prior investment in large Arctic field programs (e.g., SHEBA, LAII/ATLAS, ARM) that provide information for testing improved high-latitude physics and parameterizations. The ASR will be based on the polar-optimized version of the Weather Research and Forecasting (WRF) model and the Arctic enhancements developed for this project will be funneled through the National Center for Atmospheric Research for release to the scientific community. Following the precedent set by the NOAA Climate Diagnostics Center in its access and visualization tools developed for the NCEP/NCAR reanalysis, we will provide similar access to ASR output to ensure its widespread use. We will also design educational modules to introduce the media and broader public to the ASR and its applications. We will use our university affiliations to entrain graduate and undergraduate students into the testing and uses of ASR output. Finally, we will work with Arctic indigenous people to merge ASR with their experiences to evaluate the "strange" Arctic weather.

The ASR will serve to synthesize IPY observations, and help shape the legacy of the IPY Arctic observing network through providing an ideal platform for observing system sensitivity studies as recommended by the NRC. In addition to CARE/ASR, ASR has considerable overlap with THORPEX-IPY (full proposal 121) that will accelerate improvements in the prediction and understanding of high-impact Arctic weather on the 1 to 14-day time-scale. A strong link is established with the European Centre for Medium-Range Weather Forecasts (ECMWF).

The ASR is supported through start-up funding supplied by NOAA, and through NSF's IPY support.

General Strategy for the ASR:

The first generation ASR will span the years 2000-2010, encompassing the years since the launch of the NASA EOS spacecraft. With the framework in place, further efforts should provide systematic updates and extend the reanalysis to earlier years, eventually to 1957 (the International Geophysical Year).

The ASR system will generate fields at high spatial (10 km) and temporal (3 h) resolution. The domain will include the Arctic Ocean and its terrestrial drainage. Seventy one levels in the vertical are planned with a model top at 10 hPa. The ASR will build upon lessons learned from existing reanalyses by optimizing model physics and parameterizations to Arctic conditions, as well as by optimizing methods of data assimilation.

A fundamental issue in the ASR design is the choice of system components to be coupled in the assimilation process. On the one hand, the coupling of components allows for the inclusion of interactions and feedbacks. On the other hand, the varying levels of maturity of the component modules introduce the risk of biases that can contaminate the coupling and the final product. This issue has been addressed in various workshops (SEARCH Workshop of November 2001; SEARCH Open Science Meeting, 2004; the April 2006 workshop noted above). The emerging conclusion is that atmosphere and land surface models are sufficiently advanced to merit inclusion in the ASR. Sea ice conditions will be specified in this first ASR version to avoid the much higher level of complexity arising from the coupling of the atmosphere and ocean.

ASR Plan and Schedule:

Development of the ASR will demand tight collaboration between the four participating institutions. The lead institution will be the Polar Meteorology Group (PMG) of Byrd Polar Research Center (BPRC) at The Ohio State University. David Bromwich of the BPRC PMG will lead in the development of the atmospheric parameterizations for the WRF model and will develop the sea ice model to be used in the Noah surface package. The BPRC PMG will perform the first generation ASR for 2000-2010. Efforts led by Zhiquan Liu at the Mesoscale and Microscale Meteorology Division of the National Center for Atmospheric Research (NCAR, MMM) will optimize the WRF-Var system for data assimilation of Arctic observations, and verify that the WRF-Var system will correctly assimilate all data streams. Bill Kuo will lead the effort to supply processed, quality controlled datasets produced by UCAR's COSMIC program. Fei Chen in conjunction with Michael Barlage of NCAR will optimize the Noah-based High Resolution Land Data Assimilation System for use over Arctic land areas. John Walsh of the University of Illinois (UIUC) will have primary responsibility for (1) organization and evaluation (including some quality control) of the various input data streams for the ASR, (2) OSE work to address the observing system requirements (3) data archiving, and developing data access and visualization tools (overseeing the work of W. Chapman). Efforts by Mark Serreze and Andrew Slater of the Cooperative Institute for Research in Environmental Sciences at the University of Colorado-Boulder (CIRES/CU) will focus on (1) optimizing the Noah land surface model components and land surface data assimilation in collaboration with NCAR, (2) evaluation of the ASR products (following on the Serreze group's evaluations of the Arctic performance of global reanalyses), (3) working with UIUC to test and refine web tools to assure an optimal, user-friendly configuration.

The ASR will be performed on the Ohio Supercomputer Center's (OSC) computers by the BPRC PMG that has extensive experience performing multi-decadal and multi-terrabyte regional simulations in the polar regions. All observations must be transferred to OSC computers, including the Binary Universal Format Representation (BUFR) archive of the operational NCEP stream. The reanalysis is estimated to produce 5TB of raw output per year. This will be processed to perhaps 10% of this volume for users, following the general approach of the North American Regional Reanalysis.

The timetable can be summarized as follows. Year 1 will be devoted to testing and refinement of models, the data assimilation system, and data streams. During Year 2, we will merge these components and run short test assimilations. Year 3 will focus on running the ASR for 3 years to verify assimilation of all data streams, and finalizing the data archival and distribution system. Efforts during year 4 will complete the reanalysis for the period 2000-2010 at low resolution (ASR-Interim).

ASR Presentations:

ASR Publications:

Links