D.L. Martin and K.C. Partington
National / Naval Ice Center, 4251 Suitland Rd., Washington, D.C. 20395, U.S.A.
Introduction
The mandate of the National / Naval Ice Center is to provide sea-ice analyses, forecasts, outlooks and ship-routing recommendations within the marginal ice zone of all Arctic and Antarctic seas, with support from U.S. Navy, NOAA and the U.S. Coastguard. A necessary and critical element of that support is the PIPS coupled-ice ocean model maintained by FNMOC. This model supports the forecasting function required of NIC and the accuracy of these products has a major impact on the quality of NIC's support. With its direct contact with national customers, enhanced recently through a customer survey, and the availability of a near-real time stream of sea-ice observations from many sources, the NIC is in a unique position to evaluate the operational requirements for PIPS 3.0. This short document summarizes these requirements in order to provide a framework for PIPS 3.0 developments.
PIPS 3.0 Requirements
The overall requirement of PIPS 3.0, from an operational perspective, is to improve the
accuracy of the existing forecast products and to improve the spatial resolution of the model
to the point where fracture mechanics and ambient noise can be resolved (in a statistical
sense). PIPS 3.0 should play a major part in enabling the NIC to move resources away
from (manual) global ice chart production and towards the production of daily high spatial
resolution regional ice analyses, in line with customer requirements.
One way in which the quality of the PIPS model forecasts can be improved is by updating the ocean model, including the provision of better vertical resolution, improved bathymetric spatial resolution and replacement of the Bryan-Cox primitive equation by a free surface formulation. However, the over-riding way in which the quality issue must be addressed is by improved data assimilation, replacing the climatological ocean model by one that assimilates observations and by improving the range and quality of assimilated ice observations. Data assimilation in PIPS 2.0 is extremely primitive. An example of this is provided by the ingestion of ice concentrations from SSM/I, which are subject to gross errors under certain conditions and yet are effectively assimilated into PIPS 2.0 under an assumption of "zero error" (i.e. the model physics is discarded routinely in favor of these flawed observations). Other data, for example ice drift vectors from satellites which are derived using mature algorithms, are not used at all. A major operational requirement of PIPS 3.0 is therefore to improve the range and quality of the observations used to initialize the model and to improve the method of assimilation. The specific types of observations which should be focussed on initially, from an operational perspective, include ice concentration, ice drift and ocean currents. Assimilation of these observations presents some major challenges. Of particular note is the problem of assimilating ice drift vectors which represent ice movement within a very narrow band of temporal frequencies and it will be important to ensure that the assimilation procedure is carried out in a physically meaningful manner.
The second area of operational improvement of PIPS 3.0 lies in its improved spatial
resolution to 10km or better, to the point where it could support fracture mechanics, by use
of an anisotropic elastic-viscous-plastic rheology and hence ambient noise prediction. The
development in physics required for this will form a key component of the PIPS 3.0
program, but the overall result will be to improve the range of NIC products to state of the
art. Assessment of these products will be a challenge which NIC, through its operational
customers and sponsors, as well as its science program, will be keen to support.
Recommendations for the PIPS 3.0 Program
The operational requirements for PIPS 3.0 described above lead to a set of
recommendations for the PIPS 3.0 program. These are listed as follows:
€ The development of PIPS 3.0 should be carried out within the context of a data
assimilation strategy in which developments in physics are matched by developments in
data assimilation techniques (Partington and Steffen, 1998). The balance is important
not only to ensure that the model physics is adequately constrained, but also to ensure
that observations are used in a planned, coherent and optimized manner. Such a
balanced development would also pave the way for a high quality Arctic-wide now-cast
product which would further serve NIC requirements by allowing NIC to focus manual
(ice analyst) support in areas of national interest.
€ The NIC, having near real time data sources and a new science program, is in a
position to advise on operational requirements and to help evaluate model developments
(either through providing test datasets and/or through themselves undertaking
evaluations of forecasting performance). NIC would therefore like to encourage
collaborations, particularly at the product definition and product evaluation phases of
projects.
Reference
Partington, K.C. and Steffen, K., 1998, "Proposed Development of a Joint
Scientific-Operational Arctic-Wide Sea-Ice Product", unpublished white paper prepared for
NASA, 14 April 1998.