Figure 1.
The usual coverage of conventional observations inside the AROME-Arctic domain (example taken for 4 August 2012, 12UTC) (a) and the number of observations available in ECMWF database (b).
Figure 1.
The usual coverage of conventional observations inside the AROME-Arctic domain (example taken for 4 August 2012, 12UTC) (a) and the number of observations available in ECMWF database (b).
Figure 2.
The absolute (top) and relative (bottom) degrees of freedom for signal (DFS) expressing the sensitivity of the assimilation system to different observed parameters in use, where IASI + ATOVS—conventional data + IASI + ATOVS; ATOVS—conventional data + AMSU-A + AMSU-B/MHS; AMSUA—conventional data + AMSU-A; and AMSUB—conventional data + AMSU-B/MHS.
Figure 2.
The absolute (top) and relative (bottom) degrees of freedom for signal (DFS) expressing the sensitivity of the assimilation system to different observed parameters in use, where IASI + ATOVS—conventional data + IASI + ATOVS; ATOVS—conventional data + AMSU-A + AMSU-B/MHS; AMSUA—conventional data + AMSU-A; and AMSUB—conventional data + AMSU-B/MHS.
Figure 3.
The coloured dots show the positions of the available verifying (surface and radiosonde) stations inside the AROME-Arctic model.
Figure 3.
The coloured dots show the positions of the available verifying (surface and radiosonde) stations inside the AROME-Arctic model.
Figure 4.
The impact of surface data assimilation on the temperature (a), geopotential (b), 10-m wind (c), and 2-m temperature (d). While for (a,b), positive/negative values shown a positive/negative impact, for (c), the higher the skill the better the impact and for (c,d) the green and red lines show, respectively, the experiment with and without surface data assimilation. RMSE, STDV, and BIAS, respectively, stand for root-mean-square error, error standard deviation, and bias.
Figure 4.
The impact of surface data assimilation on the temperature (a), geopotential (b), 10-m wind (c), and 2-m temperature (d). While for (a,b), positive/negative values shown a positive/negative impact, for (c), the higher the skill the better the impact and for (c,d) the green and red lines show, respectively, the experiment with and without surface data assimilation. RMSE, STDV, and BIAS, respectively, stand for root-mean-square error, error standard deviation, and bias.
Figure 5.
The verification of the analyses and forecasts of dew point temperature against observations expressed as mean scores over different model lead times (a) and score at the 700 hPa model level (b,c). Please refer to the legends for the different plots. STDV, BIAS, OBS, and CASES, respectively, stand for error standard deviation, bias, observation, and number of used cases in the verification.
Figure 5.
The verification of the analyses and forecasts of dew point temperature against observations expressed as mean scores over different model lead times (a) and score at the 700 hPa model level (b,c). Please refer to the legends for the different plots. STDV, BIAS, OBS, and CASES, respectively, stand for error standard deviation, bias, observation, and number of used cases in the verification.
Figure 6.
The relative root-mean-square error (RMSE) change in the forecast of relative humidity when adding AMSU-B/MHS (a), IASI (c), and AMSU-A (d) in the data assimilation system. The positive/negative values show positive/negative impacts of the satellite instruments. The graph in (b) shows the significant test applied to the normalized mean RMSE difference to the case at 850 hPa.
Figure 6.
The relative root-mean-square error (RMSE) change in the forecast of relative humidity when adding AMSU-B/MHS (a), IASI (c), and AMSU-A (d) in the data assimilation system. The positive/negative values show positive/negative impacts of the satellite instruments. The graph in (b) shows the significant test applied to the normalized mean RMSE difference to the case at 850 hPa.
Figure 7.
The verification against the AMSU-B/MHS channel 5 brightness temperature (a,b) and the verification against radiosonde observations (c). The horizontal axes in (a,b) show forecast lengths similar to the one in (c).
Figure 7.
The verification against the AMSU-B/MHS channel 5 brightness temperature (a,b) and the verification against radiosonde observations (c). The horizontal axes in (a,b) show forecast lengths similar to the one in (c).
Figure 8.
The AROME-Arctic domain and the subdomains for Moist Total Energy Norm computation.
Figure 8.
The AROME-Arctic domain and the subdomains for Moist Total Energy Norm computation.
Figure 9.
The analysis and forecast from 12 UTC, 6 December 2013. There is a polar low developing near Novaya Zemlya in these forecasts.
Figure 9.
The analysis and forecast from 12 UTC, 6 December 2013. There is a polar low developing near Novaya Zemlya in these forecasts.
Figure 10.
The analysis and forecast from 00 UTC, 10 December 2013. The dominating atmospheric systems are the decaying polar low near Novaya Zemlya and the developing synoptic scale cyclone moving relatively fast throughout the model domain.
Figure 10.
The analysis and forecast from 00 UTC, 10 December 2013. The dominating atmospheric systems are the decaying polar low near Novaya Zemlya and the developing synoptic scale cyclone moving relatively fast throughout the model domain.
Figure 11.
The analysis and forecast from 12 UTC, 15 December 2013. These forecasts are dominated mainly by large scale and stationary large gradient pressure system.
Figure 11.
The analysis and forecast from 12 UTC, 15 December 2013. These forecasts are dominated mainly by large scale and stationary large gradient pressure system.
Figure 12.
The analysis and forecast from 00 UTC, 19 December 2013. The forecasts show stationary atmospheric phenomena.
Figure 12.
The analysis and forecast from 00 UTC, 19 December 2013. The forecasts show stationary atmospheric phenomena.
Figure 16.
A vertical cross section of the relative humidity (coloured pattern) and a normal-wind field (black lines) along the line shown in the right hand side map. The larger circle shows the upper-tropospheric synoptic-scale cyclone, and the small one shows the polar low (acting below roughly 800 hPa). The cross section of humidity shows a dry air at the centre of the low as a signature of the stratospheric air intrusion, as found during the campaign observation (Linders and S�tra [
33]; Kristj�nsson et al. [
10]. The plots are using a 24-h forecast, valid at 12 UTC 8 December 2013, from the run with all observations (ARCIASI).
Figure 16.
A vertical cross section of the relative humidity (coloured pattern) and a normal-wind field (black lines) along the line shown in the right hand side map. The larger circle shows the upper-tropospheric synoptic-scale cyclone, and the small one shows the polar low (acting below roughly 800 hPa). The cross section of humidity shows a dry air at the centre of the low as a signature of the stratospheric air intrusion, as found during the campaign observation (Linders and S�tra [
33]; Kristj�nsson et al. [
10]. The plots are using a 24-h forecast, valid at 12 UTC 8 December 2013, from the run with all observations (ARCIASI).
Figure 17.
The development of the polar low (pointed with red arrows) between 18 UTC, 6 December 2013 and 12 UTC, 8 December 2013 through different analyses.
Figure 17.
The development of the polar low (pointed with red arrows) between 18 UTC, 6 December 2013 and 12 UTC, 8 December 2013 through different analyses.
Figure 18.
The development of the polar low (pointed with red arrows) between 00 UTC, 9 December 2013 and 00 UTC, 11 December 2013 through different analyses.
Figure 18.
The development of the polar low (pointed with red arrows) between 00 UTC, 9 December 2013 and 00 UTC, 11 December 2013 through different analyses.
Figure 19.
The position of the lowest points (shown next to the experiment names) of the forecasted low projected on the forecast using all observation (ARCIASI, 986 hPa). Inside the AROME-Arctic domain, Franz Josef land, Novaya Zemlya, Swalbard, and northern Norway can be seen. The different legends are as follows: not filled circle (ARCAMSUB, 986 hPa), not filled square (ARCATOVS, 990 hPa), plus-sign (ARCAMSUA, 984 hPa), blue triangle (ARCCONV, 984 hPa), and diamond (ARCREF, 986 hPa). Note that ARCIASI and ARCAMSUB predict very similar lows for both the position and intensity.
Figure 19.
The position of the lowest points (shown next to the experiment names) of the forecasted low projected on the forecast using all observation (ARCIASI, 986 hPa). Inside the AROME-Arctic domain, Franz Josef land, Novaya Zemlya, Swalbard, and northern Norway can be seen. The different legends are as follows: not filled circle (ARCAMSUB, 986 hPa), not filled square (ARCATOVS, 990 hPa), plus-sign (ARCAMSUA, 984 hPa), blue triangle (ARCCONV, 984 hPa), and diamond (ARCREF, 986 hPa). Note that ARCIASI and ARCAMSUB predict very similar lows for both the position and intensity.
Figure 20.
A vertical cross section of the 24-h forecasts valid at 12 UTC on 8 December 2013 issued from different experiments. The unit of wind is in m/s. The cross-sectional lines were chosen differently to better check the near surface vortice.
Figure 20.
A vertical cross section of the 24-h forecasts valid at 12 UTC on 8 December 2013 issued from different experiments. The unit of wind is in m/s. The cross-sectional lines were chosen differently to better check the near surface vortice.
Table 1.
The use of observations in the AROME-Arctic. Note, 10-m winds are assimilated over sea�only.
Table 1.
The use of observations in the AROME-Arctic. Note, 10-m winds are assimilated over sea�only.
Type | Parameter (Channel) | Bias Correction | Thinning |
---|
TEMP | U, V, T, Q | No | No |
SYNOP | Z, V10m, U10m | No | Temporal and spatial |
DRIBU | Z | No | Temporal and spatial |
AIREP | U, V, T | No | 25 km horizontal |
AMSU-A | Channels (see Table 2a) | Variational | 80 km horizontal |
AMSU-B, MHS | Channels (see Table 2b) | Variational | 80 km horizontal |
IASI | Channels (see Table 2c) | Variational | 80 km horizontal |
Table 2.
(a) The conditions for the assimilation of advanced microwave sounding unit-A (AMSU-A) channels in AROME-Arctic. Each condition is necessary but not sufficient. The�obs-fg is the observation minus the simulated radiance in observation space; (b) the conditions for assimilation of advanced microwave sounding unit-B (AMSU-B)/humidity-sensitive microwave (MHS) channels in AROME-Arctic. Each condition is necessary but not sufficient. The�obs-fg is the observation minus the simulated radiance in observation space; and (c) the conditions for the assimilation of Infrared Atmospheric Sounding Interferometer (IASI) channels in AROME-Arctic.
(a)
Assimilation Conditions | AMSU-A Channels |
---|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 |
---|
3 < scan position < 28 | | | | | | x | x | x | x | x | | | | | |
Over open sea | | | | | | x | x | x | x | x | | | | | |
Over sea ice | | | | | | | x | x | x | x | | | | | |
Over land | | | | | | | x | x | x | x | | | | | |
Clear |obs-fg| ch 4 ≤ 0.7 K | | | | | | x | x | x | x | x | | | | | |
Cloudy |obs-fg| ch 4 > 0.7 K | | | | | | | | x | x | x | | | | | |
(b)
Assimilation Conditions | AMSU-B/MHS Channels |
---|
1 | 2 | 3 | 4 | 5 |
---|
9 < scan position < 82 | | | x | x | x |
Over open sea and |obs-fg| ≤ 5 K | | | x | x | x |
Over land and |obs-fg| ≤ 5 K and | | | x | x | |
model orography < 1000/1500 m | | | | | |
for channels 3; 4 | | | | | |
(c)
Assimilation Conditions
| IASI Channels
|
---|
Over open sea | 38, 51, 63, 85, 104, 109, 167, 173, 180, |
| 185, 193, 199, 205, 207, 212, 224, 230, |
| 236, 239, 242, 243, 249, 252, 265, 275, |
| 294, 296, 306, 333, 337, 345, 352, 386, |
| 389, 432, 2919, 3008, 3014, 3069, 3087, |
| 3098, 3207, 3228, 3281, 3309, 3322, 3339, |
| 3438, 3442, 3484, 3491, 3499, 3506, 3575, |
| 3582, 3658 |
Over land | 38, 51, 63, 85, 104, 109, 167, 173, 180, |
| 185, 193, 199, 205, 207, 212, 224, 230, |
| 236, 239, 242, 243, 249, 252, 265, 275, |
| 294, 296, 306, 345, 386, 389, 432, 2919, |
| 3069, 3087, 3098, 3281, 3309, 3339, 3442, |
| 3484, 3491, 3499, 3506, 3575, 3582, 3658, |
| 4032 |
Over sea ice | 51, 63, 85, 87, 104, 109, 167, 173, 180, |
| 185, 193, 199, 205, 207, 212, 224, 239, |
| 265, 275, 294, 306, 2701, 2819, 2910, |
| 2991, 2993, 3002, 3008, 3014, 3027 |
Table 3.
The number of active observations for data assimilation for 7 December 2013.
Table 3.
The number of active observations for data assimilation for 7 December 2013.
| 00UTC | 03UTC | 06UTC | 09UTC | 12UTC | 15UTC | 18UTC | 21UTC |
---|
Surf. Pressure, land | 25 | 18 | 34 | 19 | 31 | 19 | 32 | 17 |
Surf. Pressure, auto | 60 | 60 | 51 | 60 | 55 | 59 | 53 | 60 |
Surf. Pressure, ship | 2 | | 4 | | 3 | | 3 | |
Surf. Wind, ship | 4 | | 8 | | 6 | | 10 | |
AMDAR Temperature | 15 | 22 | 16 | 54 | 92 | 52 | 41 | 28 |
AMDAR Wind | 64 | 42 | 32 | 108 | 184 | 102 | 80 | 56 |
Dribu Pressure | 10 | 10 | 10 | 10 | 10 | 10 | 10 | 10 |
Radiosonde Wind | 422 | | 138 | | 642 | | 144 | |
Radiosonde Temperature | 132 | | 33 | | 185 | | 27 | |
Radiosonde Humidity | 67 | | 20 | | 94 | | 18 | |
METOP-A AMSU-A | | | 146 | 373 | 561 | 1104 | 1061 | 499 |
METOP-A MHS | | | 142 | 1042 | 832 | 1015 | 809 | 377 |
METOP-A IASI | | 15 | 6752 | 18,628 | 16,383 | 16,957 | 12,752 | 8157 |
NOAA-15 AMSU-A | 9 | 346 | 465 | 349 | 802 | 174 | | |
NOAA-18 AMSU-A | 187 | 630 | 395 | 316 | 302 | 608 | | |
NOAA-18 MHS | 79 | 884 | 982 | 210 | 308 | 147 | | |
NOAA-19 AMSU-A | 850 | 566 | 501 | 765 | 224 | | | 157 |