Evaluation of Reanalysis Products in the Arctic

How well can we estimate the state of the atmosphere and the rate it is changing in the polar regions in retrospective analyses? In the data sparse Arctic, atmospheric analyses are poorly constrained by observations and are strongly influenced by model parameterizations. There are currently seven different sets of global reanalysis products that are current or near current in temporal coverage: NCEP-1, NCEP-2, CFSR, 20CR, MERRA, ERA-Interim, and JRA-25 (definitions follow).

Retrospective analyses have been a critical tool in studying weather and climate variability for the last 15 years. Reanalyses blend the continuity and breadth of output data from a numerical model with the constraint of vast quantities of surface, radiosonde, and satellite observational data. The result is a long-term continuous and spatially complete data record. Reanalysis products are used in many different applications including evaluation of atmospheric circulation patterns and processes, change detection, the forcing of ice–ocean models, regional atmospheric models, land models, or air chemistry models, and for establishing the initial conditions for forecast models. Better understanding the strengths and weaknesses of these products will improve our ability to evaluate the long-term trends in the rapidly changing Arctic environment and may also improve our ability to make seasonal projections of sea ice and weather conditions in the Arctic.

In this focused study we will compare the monthly mean estimates of the surface and tropospheric air temperature, the surface pressure and winds, the total precipitation, and the surface and top-of-the-atmosphere radiative fluxes in a three-tiered set of analyses. The first-order comparisons will be made to independent point observations from a selected set of land stations and drifting ice stations. The second-order comparisons will be made of the statistical properties (mean, standard deviation, and extremes) of the fields from each of the reanalyses. Finally, the third-order comparisons will be made of the 30-year trends in the fields of each of the reanalyses. While this is an ambitious project to accomplish in just one year, we plan to use the economies of scale to perform the identical analysis procedures on all six of the reanalysis products.

The ultimate goal is to better understand the strengths and weaknesses of these products in a data-sparse region where the reanalysis models may differ the most. Better understanding of these products may improve the ability for NOAA to make seasonal projections of sea ice and weather conditions in the Arctic.