Geoscience Reference
In-Depth Information
(a)
(b)
0.03
0.08
0.06
0.02
0.04
0.01
0.02
0
0
-0.01
-0.02
-0.02
-0.04
-0.03
-0.06
0
1
2
3
4
5
6
7
8
9
10
11
0
1
2
3
4
5
6
7
8
9
10
11
Forecast day
Forecast day
Fig. 2 Normalised root mean square forecast error difference for the ECMWF 1,000 hPa geopotential for
(a) Cressman minus OI snow analyses, both using the SYNOP reports and the 24-km NOAA NESDIS IMS
snow cover product and (b) Cressman minus OI, with OI using SYNOP and 4-km NOAA NESDIS IMS
snow cover data (new ECMWF snow analysis) and Cressman using SYNOP and the 24-km IMS product
(old ECMWF snow analysis). Statistics are computed based on daily analyses at 00 UTC from 01 December
2009 to 28 February 2010. Vertical bars show the 90 % confidence interval. The x-axis shows the forecast
range from 0 to 10 days. Positive impact of the OI analysis compared to the Cressman analysis is shown by
positive values
SYNOP snow depth data than Cressman. The difference between the two analyses mainly
results from differences in the structure functions between OI and Cressman.
The 2009/2010 winter season, with cold and snowy conditions in the northern hemisphere
(Cohen et al.
2010
), highlighted the importance of good-quality snow analysis (de Rosnay
et al.
2011b
). Figure
2
shows the impact of different snow analysis configurations on the
forecast 1000 hPa geopotential height error for the winter 2009-2010. Figure
2
a shows the
impact of the OI snow analysis compared to the Cressman snow analysis, with both schemes
using SYNOP and IMS snow cover data at 24 km resolution. Figure
2
b shows the impact of
the revised ECMWF analysis implemented in 2010 compared to the previous analysis. The
old analysis uses Cressman and observations from the SYNOP network and the IMS 24 km
product. The new analysis relies on the OI and uses observations from SYNOP and the 4-km
snow cover IMS product. The new analysis also accounts for an improved pre-processing and
quality control of the IMS NESDIS data. In particular, based on an altitude threshold of
1500 m, the use of the IMS data is switched off in mountainous areas. Replacing the
Cressman snow analysis by the OI has a relatively neutral impact on the atmospheric cir-
culation, although a slight non-significant improvement can be seen (Fig.
2
a). The new
ECMWF snow analysis (Fig.
2
b) has an overall positive impact on the atmospheric forecasts
skill, with root mean square error forecast for the 1000 hPa geopotential height improved by
1-4 % in the short range (forecasts until day 4). Figures
1
and
2
illustrate that the combined
improvements of the analysis approach (OI vs Cressman) and data pre-processing and quality
control (IMS snow cover product resolution and altitude threshold) lead to improve both the
snow depth fields and the low-level atmospheric forecast.
3 Soil Moisture Analysis
3.1 History of Soil Moisture Analysis for NWP
In the absence of a near-real-time global network for providing soil moisture information,
using screen-level data has been the only source of information that has been continuously
