Advancing Antarctic Science with AMPS

by Rachel Hauser

 Having advance notice of poor weather is useful to anyone, but it is particularly critical for those working in remote polar regions where weather can quickly amplify existing hazardous conditions. To maximize its research mission safety and effectiveness, the National Science Foundation’s Office of Polar Programs supports the Antarctic Mesoscale Prediction System (AMPS), which aids weather prediction in Antarctica. Working with The Ohio State University (OSU) and University of Colorado, members of NCAR’s Mesoscale and Microscale Meteorology Division have developed a real-time, experimental Weather Research and Forecasting (WRF) model-based system that provides numerical weather prediction guidance for U.S. Antarctic Program (USAP) forecasters. While the system primarily serves the USAP, it also offers support to scientific field campaigns as well as a host of nations operating research stations and experiments in Antarctica. Using a polar-modified version of WRF, AMPS delivers the highest resolution model output available on the continent.

A need for a better, near-real-time weather forecasting capability came to the fore when Jerri Nielsen, a physician working at the U.S. research base at South Pole in 1999, was diagnosed with breast cancer during the austral winter, when weather conditions make flying to the pole impossible. The forecast challenge was to evacuate Nielsen as early as possible.

“At a subsequent National Science Foundation Antarctic weather forecasting workshop, stimulated in part by the Nielsen incident, it was noted that personnel safety and research would benefit from more tailored numerical weather prediction capabilities,” recounts David Bromwich, a professor of atmospheric science at The Ohio State University.

Timing to create a near-real-time model was ideal as a polar-modified version of MM5, developed by OSU, had been shown to perform skillfully over Greenland. The Polar MM5 was the weather model used for AMPS weather prediction until 2008 when the MM5 reached the end of its lifespan, says Bromwich. The polar modifications developed for the MM5 had been ported to NCAR’s Advanced Research WRF (ARW) model, which was then exclusively used in AMPS. Today, the ARW provides AMPS with numerical weather prediction capabilities that are used for a range of scientific and logistical activities in Antarctica. Since the original work began, many of the polar modifications that capture features unique to high latitudes and extensive ice sheets have been included in the officially-released versions of WRF. Among the modifications are provision of fractional sea ice coverage in grid cells and better representations of ice sheet conditions, both of which influence polar weather.  

AMPS forecast domains. Outer frame: 15-km spacing. Inner frames: 5-km spacing. Innermost frame (Ross Island area): 1.67-km spacing.

ARW Antarctic weather forecasts are generated on grids with horizontal spacings that range from 1.7 to 15 km. The lower resolution coverage occurs in areas where USAP research and administrative requirements are limited; these areas lie over the ocean and around the continent, as far north as Australia and southern South America, while the highest resolution data are generated around McMurdo Station, the USAP’s main area of operation. WRF, says Jordan Powers, an NCAR scientist and project lead for AMPS, is the highest resolution weather model covering the Antarctic, with resolution of the key continental regions provided at 5 km or better.

AMPS furnishes twice-daily, high-resolution numerical guidance for Antarctica and the sub-Antarctic. Assisting in assuring model performance, USAP forecasters provide regular feedback that allows the AMPS team to upgrade the system on an ongoing basis. Beyond AMPS’s use for daily forecasting, special applications of it have figured prominently, including support for scientific field campaigns, medical and marine rescues, as well as assisting, as needed, with the activities of other nations operating in Antarctica. In addition, AMPS provides a capability for process and event studies, a platform to test model parameterizations in a polar region, and a database of high-resolution numerical output to researchers working on Antarctic meteorology and climatology. The latter is the AMPS archive, which contains model output since 2001. Notably, the archived short-term forecasts are being used to understand and predict possible regional weather conditions throughout the year. This information is also useful for exploring climatic conditions over the continent, says Bromwich.
Recent AMPS efforts include implementing a separate forecast domain of 2.5-km resolution over the Antarctic Peninsula in austral summer 2009-2010 for the LARISSA (Larsen Ice Shelf System, Antarctica) field campaign. Funded by NSF, LARISSA is an interdisciplinary program investigating the structure and possible changes to the Larsen Ice Shelf, on the western side of the Antarctic Peninsula. A better understanding of local small-scale processes will assist scientists in refining models that estimate how larger parts of the Antarctic cryosphere might respond to environmental change.

“Part of the AMPS effort includes providing tailored support for field campaigns, such as we did with LARISSA,” Powers explains. “For the 2010-2011 field season we are providing a special, 5-km forecast grid to serve the large number of research efforts taking place along the Central Transantarctic Mountains.”  This is another example of AMPS’s service toward NSF’s Antarctic goals.


AMPS 24-h forecast of surface temperatures (deg C) for the CTAM (Central TransAntarctic Mtns.) forecast grid (5-km spacing) set up for the USAP for the 2010-2011 field season. Valid: 0000 UTC 16 October 2010.