An Air Mass Transformation Model for Short-Range Weather Forecasting

J. Reiff Royal Netherlands Meteorological Institute (KNMI), 3730 AE De Bilt, The Netherlands

Search for other papers by J. Reiff in
Current site
Google Scholar
PubMed
Close
,
D. Blaauboer Royal Netherlands Meteorological Institute (KNMI), 3730 AE De Bilt, The Netherlands

Search for other papers by D. Blaauboer in
Current site
Google Scholar
PubMed
Close
,
H. A. R. De Bruin Royal Netherlands Meteorological Institute (KNMI), 3730 AE De Bilt, The Netherlands

Search for other papers by H. A. R. De Bruin in
Current site
Google Scholar
PubMed
Close
,
A. P. Van Ulden Royal Netherlands Meteorological Institute (KNMI), 3730 AE De Bilt, The Netherlands

Search for other papers by A. P. Van Ulden in
Current site
Google Scholar
PubMed
Close
, and
G. Cats Royal Netherlands Meteorological Institute (KNMI), 3730 AE De Bilt, The Netherlands

Search for other papers by G. Cats in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A model has been constructed for the purpose of forecasting at specific places the potential temperature, the specific humidity and the depth of the boundary layer. The model, an Air Mass Transformation model (AMT model), consists of a trajectory model and a one-dimensional boundary layer model. The horizontal and vertical advection of air masses is described by trajectories. The trajectories are computed from the analyzed geostrophic wind fields of a large-scale grid-point model. The lowest trajectory advects the air in the boundary layer. The air-mass transformations in the boundary layer are described by a simple boundary layer model. During unstable conditions, a slab model is used; during stable conditions, linear potential temperature profiles and humidity profiles are assumed. Along the lowest trajectory, the surface fluxes of heat and water vapor are computed from a simple parameterization scheme. The initial conditions for the state of the boundary layer are obtained from an analysis scheme. In this scheme, the routinely collected radiosonde data in the source area are used. So far the model has been tested for the period July–September 1981. During this period, twice-daily 12-hour model runs have been made. DeBilt (52°N, 5°E) was chosen as the test site. The following results have been obtained:

  1. The model temperature “forecasts” show a 0.85 correlation coefficient with the observed values at noon and 0.80 at midnight.

  2. The model has a strong potential to forecast the occurrence of boundary layer clouds.

  3. A vertical resolution of about 50 mb in the analysis scheme is necessary to obtain these results.

  4. Specific humidity forecasts based on pure (12-hour) advection show correlation coefficients of 0.81 and 0.77 with the observed humidity at noon and midnight respectively.

Abstract

A model has been constructed for the purpose of forecasting at specific places the potential temperature, the specific humidity and the depth of the boundary layer. The model, an Air Mass Transformation model (AMT model), consists of a trajectory model and a one-dimensional boundary layer model. The horizontal and vertical advection of air masses is described by trajectories. The trajectories are computed from the analyzed geostrophic wind fields of a large-scale grid-point model. The lowest trajectory advects the air in the boundary layer. The air-mass transformations in the boundary layer are described by a simple boundary layer model. During unstable conditions, a slab model is used; during stable conditions, linear potential temperature profiles and humidity profiles are assumed. Along the lowest trajectory, the surface fluxes of heat and water vapor are computed from a simple parameterization scheme. The initial conditions for the state of the boundary layer are obtained from an analysis scheme. In this scheme, the routinely collected radiosonde data in the source area are used. So far the model has been tested for the period July–September 1981. During this period, twice-daily 12-hour model runs have been made. DeBilt (52°N, 5°E) was chosen as the test site. The following results have been obtained:

  1. The model temperature “forecasts” show a 0.85 correlation coefficient with the observed values at noon and 0.80 at midnight.

  2. The model has a strong potential to forecast the occurrence of boundary layer clouds.

  3. A vertical resolution of about 50 mb in the analysis scheme is necessary to obtain these results.

  4. Specific humidity forecasts based on pure (12-hour) advection show correlation coefficients of 0.81 and 0.77 with the observed humidity at noon and midnight respectively.

Save