Studies on the Impacts of 3D-VAR Assimilation of Satellite Observations on the Simulation of Monsoon Depressions over India - Data Assimilation for Atmospheric, Oceanic and Hydrologic Applications - page 658

assimilation experiments except for the MODIS run. The ATOVS run shows higher
error values of temperature up to 850 hPa level. From 780 to 200 hPa levels the
RMSE values of temperature for the ATOVS run are found to be lower indicating
an improved forecast for temperature in the mid-troposphere and upper troposphere
for the ATOVS experiment on 20 September 2006 12 UTC. The RMSE values of
temperature decreases with height for the SSM/I run on 20 September 2006 12 UTC;
indicating a positive impact of assimilation of SSM/I total precipitable water at this
time. The ATOVS experiment shows a decrease in the RMSE values of temperature
with height for the forecast valid on 21 September 2006 12 UTC. Furthermore,
significant decrease in the error is seen above the mid-troposphere (500 hPa) level.
In contrast with the results of the ATOVS experiment, the MODIS run shows lower
RMSE values of temperature at the lower levels on the second day which then
increases with height up to 300 hPa levels, while the SSM/I run depicts higher
RMSE of temperature at all the levels, on the second day, indicating the positive
impact of assimilation of total precipitable water has somewhat decreased on the
second day of the forecast. On 22 September 2006 12 UTC, the ATOVS and the
SSM/I experiments have performed well with lower RMSE value at almost all the
levels, while the MODIS experiment shows higher RMSE values for temperature at
the lower and mid-troposphere.
RMSE of Dew-Point Temperature Profiles
Tab le 26.3 gives the domain averaged RMSE of dew-point temperature with respect
to the IMD radiosonde observation. Previous studies ( Cox et al. 1998 ; Sandeep
et al. 2006 ) have noted that the mesoscale models have difficulty in accurately
simulating the upper level moisture content. The lack of upper level moisture
observations for assimilation can be regarded as one of the reasons for the above
difficulty. Hence it is important to validate the simulated humidity profiles of all
the experiments using the available observations. As expected, the assimilation of
temperature and humidity profiles from the ATOVS and the MODIS have reduced
the RMSE values of the moisture content at the lower levels of troposphere for the
forecast valid on 20 September 2006 12 UTC when compared to the CTRL run.
However, the RMSE of dew point temperature increases with height in the middle
and upper troposphere, indicating the inability in accurately simulating the humidity
profiles even after the assimilation of observation. In contrast with the other two
assimilation experiments, the SSM/I experiment has shown significant improvement
in the simulation of humidity profiles by having lower RMSE values of the dew
point temperature for all the levels, on the day one of the forecast. The second
day of the forecast shows higher values of RMSE of the dew point temperature
at all the levels for the MODIS experiment while the ATOVS run depicts lower
RMSE values of the dew point temperature at the lower levels of troposphere. On 22
September 2006 12 UTC, both the MODIS and the ATOVS provide results having
higher RMSE values of dew-point temperature at almost all the levels while the
SSM/I run has significantly reduced the error values. Hence, it can be concluded