SLIDE 1 Evaluating representativeness errors in verification against Arctic surface
Thomas Haiden and Martin Janousek European Centre for Medium-Range Weather Forecasts
Photo IASOA
SLIDE 2 Outline
- Arctic: downward longwave radiation anomalies
- Global: 2-m temperature forecast skill
SLIDE 3
IASOA observatories
SLIDE 4 How much information about the larger Arctic area do IASOA observations contain?
IASOA
Grid-scale value at IASOA location
Larger Arctic area measurement error + representativeness synoptic-scale relationships
→ assess the spatial ‘footprint’ of IASOA observations using model analyses (ERA-Interim)
SLIDE 5
Example: Ny-Ålesund, Svalbard (79N,12E)
ERA-Interim (Δx=80 km) HRES (Δx=10 km)
SLIDE 6 Downward longwave flux at Ny-Ålesund
Grid point Land fraction r 0.62 0.918 1 0.00 0.838 2 0.18 0.816 3 0.00 0.859 67%-84% of observed variance represented Systematic and non-systematic differences between grid-points
SLIDE 7 How much information about the larger Arctic area do IASOA observations contain?
IASOA
Grid-scale value at IASOA location
Larger Arctic area representativeness synoptic-scale relationships
→ assess the spatial ‘footprint’ of IASOA observations using model analyses (ERA-Interim)
SLIDE 8 Spatial correlation of longwave flux at NyAlesund
Correlation within ERA-Interim Correlation OBS v ERA-Interim
SLIDE 9 Correlation as a function of distance
Ny-Ålesund Barrow
ERA-Interim Observation
SLIDE 10 Correlation as a function of distance
Alert Barrow
ERA-Interim Observation
SLIDE 11 Variance explained by positive correlations
Barrow, Alert, Ny-Alesund All IASOA stations
SLIDE 12 Summer (May-Oct)
Longwave flux at Barrow OBS v ERA-I
Winter (Nov-Apr)
SLIDE 13 Longwave flux at NyAlesund OBS v ERA-I
Daily Monthly
SLIDE 14
Ny-Ålesund, Svalbard (79N,12E)
ERA-Interim (Δx=80 km) HRES (Δx=10 km)
SLIDE 15 Representativeness of daily DLR (Jan 2017)
Bias Standard deviation Tiksi, Russia 3-4 W/m2 Sea-ice boundary Orography Coastal effects
SLIDE 16 Estimation based on Taylor hypothesis
~5 W/m2 (assuming 10 m/s wind speed) 16 min 2 h
SLIDE 17
2-m temperature
SLIDE 18
2-m temperature verification
NH Extratropics, 12 UTC RMSE against SYNOP
SLIDE 19
against analysis
2-m temperature verification
NH Extratropics, 12 UTC RMSE against SYNOP
SLIDE 20 2-m temperature verification
MSE NH Extratropics, 12 UTC
(2.8 K)2
against SYNOP against analysis
SLIDE 21
2-m temperature verification
MSE NH Extratropics, 12 UTC against SYNOP T850 against analysis
SLIDE 22
2-m temperature verification
MSE NH Extratropics, 12 UTC against SYNOP against analysis T850 diurnal mean
SLIDE 23
Regional variations
RMSE
SLIDE 24 Regional variations
Stations excluded where ∆z>150m
RMSE
SLIDE 25 Regional variations
SDEV
Stations excluded where ∆z>150m
SLIDE 26
Europe
SDEV
SLIDE 27 Upscaling to ~400 km (4 deg)
Small difference → larger scale issue SDEV at Day 5, 12 UTC
Problem: strong surface inversions over snow
SLIDE 28 Upscaling to ~400 km (4 deg)
Large difference → smaller scale issue SDEV at Day 5, 12 UTC
Problem: low stratus boundaries and persistence
SLIDE 29 Conclusions / applications
- Studying representativeness is a worthwhile endeavour
- Characteristics differ greatly between parameters
- Different approaches are being tested
- Scale-dependent verification (upscaling, FSS) provides insights
→ Spatial extrapolation of station observations → Assessment of ‘footprint’ of potential future obs sites → Estimation of improvements due to future resolution upgrades
SLIDE 30 Estimation based on Taylor hypothesis
Error (W/m2) 0.05*Mean (W/m2) Relative error (%)
