HIRLAM radar utilization

1
Use of radar data in the HIRLAM modelling
consortium
Magnus Lindskog Swedish Meteorological and
Hydrological Institute
Lisbon, Portugal, 8 October 2003
2
Structure

• Introduction
• Radar wind processing and assimilation
• Verification against radar derived precipitation
• Direct use of radar reflectivities
• Future plans

3
Typical weather radar
8 km 7 6 5 4 3 2 1 0

0 60 120 180
240 km
4
HIRLAM main components
assimilation
forecast
5
De-aliasing
Mapping into torus
Radial wind velocities from one scan
6
Illustration of de-aliasing
Observed velocities
De-aliased velocities
7
Filtering (Method 1)
Generation of vertical profile of horizontal
winds using different methods
Verification of different methodoligies against
radiosondes (1 Oct. 2001-1 July 2002 22610
profiles)
VAD
8
Filtering (Method 2)
Horizontal filtering
superobservations

Time filtering
unfiltered and filtered observations
filtered observations and model
9
Assimilation experiment
10 day experiment (1-10 Dec., 1999)
Three parallel runs CRLconv. obs. RWDconv.
obsrwd VADconv. obs.VAD
Integration area and radar sites
RMS of 24 h wind forecasts at 850 hPa
10
Results from pre-operational monitoring of
VAD-profiles
BIAS BG-OB wind speed (m/s)
RMS BG-OB wind speed (m/s)
11
Radar derived precipitation for verification of
products from BALTEX regional reanalysis project

12
HIRLAM Precipitation
13
Radar simulation model
CASE study 15 July, 2000 at 12 UTC
Synoptical situation
14
Radar simulation Model
15
Radar simulation Model
16
Daily ratios
computed for each grid and then
classified by altitude and month respectively.
Values from 2002.
Deviations versus altitude
Deviations versus season
17
Future plans

• Operational assimilation of VAD profiles
• Extended radial wind experiments
• Increased use of radar data for verification of
  models
• Assimilation of radar derived precipiation/reflect
  ivities
• Support operational exchange of radar data