Assessments of the reliability of NCEP circulation data and relationships with surface climate by direct comparisons with station based data

Abstract

An assessment is made of the climate simulations from the NCEP (National Centers for Environmental Prediction) Reanalyses over Europe. This assessment was initiated as part of the European Commission funded study on Atmospheric Circulation Classification and Regional Downscaling (ACCORD) and was designed to test the suitability of the Reanalyses for this type of application. Here NCEP Reanalyses (pressure, temperature and precipitation) from 1958 through 1997 are compared to station data of precipitation and temperature and composites of mean sea level pressure (MSLP) data. The comparison is made over a European window using monthly data with a focus on 3 land areas: Central and Eastern England and Italy, where daily timescale data are employed. MSLP data are generally well simulated; however, an input problem in the NCEP data prior to 1967 results in unrealistically low surface pressure. NCEP surface pressure over Greenland is also shown to be unrealistically high during the winter months. Spatially NCEP MSLP is shown to correlate quite well with UK Meteorological Office (UKMO) MSLP over the ocean and much of northeast Europe, while they correlate less well over high orographical regions. It is shown that, while daily temperature is well simulated, daily precipitation is less so, particularly during the summer months when convective precipitation is dominant. Total precipitation over the 2 UK areas is lower than observed, by as much as 22% over Central England. The number of rain day events is underestimated over the 3 regions, although the anomaly of rain per rain day is shown to vary between the regions, being overestimated in NCEP in Eastern England and Italy. Mean daily temperature is shown to be much better simulated compared with precipitation, with a slight warm bias in all 3 grid boxes.

Links between circulation and changes in the characteristics of Iberian rainfall

“…Each interpolated station series was normalized on a monthly basis, based on the 1951-80 period, and then differenced to form the index. Agreement between indices based on reanalysis and observed station data is very good (Reid et al, 2001). For winter, the correlation between the two Gibraltar minus Reykjavik indices (interpolated from NCEP and station based) is 0.99.…”

“…The only variables through time are the amount and quality of the input data. Comparison of the NCEP SLP reanalysis data for the window used here (80°to 20°N, 60°W to 70°E) with other data sets, including station data, indicates that the NCEP SLP data are reliable, particularly on the interannual time scale (Reid et al, 2001). However, data prior to 1967 should be regarded as being somewhat less reliable, particularly over Greenland (Reid et al, 2001).…”

“…However, the 40 year data period is still relatively short for studies of past variability. Other studies, for example, indicate longer-term variability in the relationships between Iberian rainfall and the NAO over the first half of the 20th century (Rodó et al, 1997;Reid et al, 2001). The changes in wet-day amount, number of rain days, and quantile amounts presented in Section 4 also raise an important question concerning the stability of the circulation-rainfall relationships.…”

“…Comparison of the NCEP SLP reanalysis data for the window used here (80°to 20°N, 60°W to 70°E) with other data sets, including station data, indicates that the NCEP SLP data are reliable, particularly on the interannual time scale (Reid et al, 2001). However, data prior to 1967 should be regarded as being somewhat less reliable, particularly over Greenland (Reid et al, 2001). For the purposes of the analyses described here (which are intended as a preliminary step in the development of downscaling methods), NCEP reanalysis SLP data (available on a 2.5°by 2.5°grid) were interpolated to a 2.5°latitude by 3.75°longitude grid (the Hadley Centre HadCM2/HadCM3 grid).…”

“…Rodó et al (1997) in the relationships between Iberian rainfall and the NAO over the first half of the 20th century (Rodó et al, 1997, see also Reid et al, 2001). 1958-97, 1958-77, and 1978-97.…”