Title: ELDAS activities at SMHIRossby Centre 2nd Progress Meeting
1ELDAS activities at SMHI/Rossby Centre 2nd
Progress Meeting
L. Phil Graham Daniel Michelson Jonas Olsson Åsa
Granström Swedish Meteorological and
Hydrological Institute, Norrköping, Sweden
2WP 2100 precipitation
3RR - 3 and 12-hour Gauge-adjusted accumulated
precipitation
- Characteristics
- 2 x 2 km horizontal resolution
- 3 and 12-hour temporal resolution
- 32-bit depth
- Uses corrected gauge observations
- 3-hour covers BALTRAD area
- 12-hour covers BALTEX Region
- Continuous production
4New Doctor!
Quality Control of Weather Radar Data for
Quantitative Application Daniel
Michelson September 2003
5Verification of Dynamic Correction Model (DCM)
with 1-hr forcing
Jokioinen, Finland
6WP 5400 Baltic demo basins
7D33) Baltic Sea Basin
D34) Swedish Basins
8 Comparison of Areal Precipitation Estimates by
Gauges and Radar A Case Study for a Central
Swedish Catchment Jonas Olsson, Barbro
Johansson, Daniel Michelson Swedish
Meteorological and Hydrological Institute SE-601
76 Norrköping, SWEDEN
9Torpshammar Basin
- Catchment
- Area 4 300 km²
- Mean annual prec. 700 mm
- Mean altitude 340 m.a.s.l. (range 20-540
m.a.s.l.)
- Data
- PTHBV interpolated gauge data
- RADAR gauge-adjusted radar data
- Daily accumulations 2000-2002 in a 4?4 km grid
(589 nodes)
10Monthly Precipitation
Torpshammar Total, 2000-2002
Differences (RADAR-PTHBV)
11Daily differences - spatial dependence (2000-2002)
(similar patterns for individual years)
12Daily differences - intensity dependence
13Runoff simulation (HBV model)
14Torpshammar Summary
- GA radar data reproduce well the gauge-observed
mean rainfall for 2002, but less accurately in
2000 and 2001. - GA radar data contain temporal and spatial
inhomogeneities, possibly related to (i) changes
in the operation and/or (ii) unstable
functionality of the instruments. - Spatial pattern of differences indicates a
consistent overestimation by the GA radar data
located north-west of the catchment (Östersund). - High mean rainfall intensities are often
underestimated and low intensities overestimated
by the GA radar data. - Catchment runoff estimated by GA radar data is
distinctly overestimated in the first half of the
period, but accurate during 2002.
(GA gauge-adjusted)
15Whats Next?
- more fair test against forecast conditions
- event-based analysis
- Glafsfjorden basin
- compare to ELDAS precipitation database
- compare soil moisture for different driving P
fields and against ELDAS generated fields - compare runoff
16more Whats Next?
- Baltic Basin runoff generation and river
discharge (WP 3200)
- European VIC application!
17European VIC application
(Variable Infiltration Capacity Model)
cooperation with University of Washington
18Why EuroVIC?
- independent source of spatially distributed soil
moisture - sets a climate perspective for SM
- additional verification (e.g. evapotranspiration)
- routed river flow over Europe from ELDAS models??
19Details of EuroVIC - 1
- ELDAS grid (0.2 x 0.2)
- climate based 1980?-2002?
- focus on ELDAS period 1999-2000
- common databases where possible (e.g. ERA-40,
ECOCLIMAP) - higher resolution orography (SRTM?)
20Details of EuroVIC - 2
- river routing included
- databases of natural river flow preferred! (who
has these?)
21old map
22European map