Sensitivity of Mesoscale Model Forecast During a Satellite Launch to Different Cumulus Parameterization Schemes in MM5 |
| |
Authors: | V Rakesh R Singh P K Pal P C Joshi |
| |
Institution: | (1) Atmospheric Sciences Division, Meteorology and Oceanography Group, Space Applications Centre (ISRO), Ahmedabad, 380015, India |
| |
Abstract: | The identification of the model discrepancy and skill is crucial when a forecast is issued. The characterization of the model
errors for different cumulus parameterization schemes (CPSs) provides more confidence on the model outputs and qualifies which
CPSs are to be used for better forecasts. Cases of good/bad skill scores can be isolated and clustered into weather systems
to identify the atmospheric structures that cause difficulties to the forecasts. The objective of this work is to study the
sensitivity of weather forecast, produced using the PSU-NCAR Mesoscale Model version 5 (MM5) during the launch of an Indian
satellite on 5th May, 2005, to the way in which convective processes are parameterized in the model. The real-time MM5 simulations
were made for providing the weather conditions near the launch station Sriharikota (SHAR). A total of 10 simulations (each
of 48 h) for the period 25th April to 04th May, 2005 over the Indian region and surrounding oceans were made using different
CPSs. The 24 h and 48 h model predicted wind, temperature and moisture fields for different CPSs, namely the Kuo, Grell, Kain-Fritsch
and Betts-Miller, are statistically evaluated by calculating parameters such as mean bias, root-mean-squares error (RMSE),
and correlation coefficients by comparison with radiosonde observation. The performance of the different CPSs, in simulating
the area of rainfall is evaluated by calculating bias scores (BSs) and equitable threat scores (ETSs). In order to compute
BSs and ETSs the model predicted rainfall is compared with Tropical Rainfall Measuring Mission (TRMM) observed rainfall. It
was observed that model simulated wind and temperature fields by all the CPSs are in reasonable agreement with that of radiosonde
observation. The RMSE of wind speed, temperature and relative humidity do not show significant differences among the four
CPSs. Temperature and relative humidity were overestimated by all the CPSs, while wind speed is underestimated, except in
the upper levels. The model predicted moisture fields by all CPSs show substantial disagreement when compared with observation.
Grell scheme outperforms the other CPSs in simulating wind speed, temperature and relative humidity, particularly in the upper
levels, which implies that representing entrainment/detrainment in the cloud column may not necessarily be a beneficial assumption
in tropical atmospheres. It is observed that MM5 overestimates the area of light precipitation, while the area of heavy precipitation
is underestimated. The least predictive skill shown by Kuo for light and moderate precipitation asserts that this scheme is
more suitable for larger grid scale (>30 km). In the predictive skill for the area of light precipitation the Betts-Miller
scheme has a clear edge over the other CPSs. The evaluation of the MM5 model for different CPSs conducted during this study
is only for a particular synoptic situation. More detailed studies however, are required to assess the forecast skill of the
CPSs for different synoptic situations. |
| |
Keywords: | Mesoscale model CPSs entrainment/detrainment skill scores TRMM synoptic validation |
本文献已被 SpringerLink 等数据库收录! |
|