1 / 65

Satellite data Monitoring

Satellite data Monitoring. Mohamed Dahoui Mohamed.Dahoui@ecmwf.int Office 160 Phone +44-118-9499 043. Outline. Introduction Why do we need monitoring? Monitoring tools Automatic alarm system Other benefits of data monitoring Conclusion. Outline. Introduction

gannon
Download Presentation

Satellite data Monitoring

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. Satellite data Monitoring Mohamed Dahoui Mohamed.Dahoui@ecmwf.int Office 160 Phone +44-118-9499 043

  2. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Other benefits of data monitoring • Conclusion

  3. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Other benefits of data monitoring • Conclusion

  4. Introduction

  5. Introduction • More than 400 million pieces of observations received daily. A lot of potential • ~12 millions are effectively used daily. More data are used indirectly (e.g. cloud detection). • Data characteristics are subject to variation • The impact of observation (on Analysis/Forecats) depends on the quality of data, spatial and temporal distribution of data, appropriate specification of errors, successful bias correction, etc.

  6. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Other benefits of data monitoring • Conclusion

  7. Why do we need monitoring • Asses the quality/availability of new data before any decision to activate the data • Help in the estimation of errors characteristics • Detects changes in the quality/availability of data • Asses the quality/availability of data during periods of blacklisting

  8. Why do we need monitoring • Monitor the impact of model/Data assimilation changes on statistics • Monitor the data impact on the forecast (FEC) • Provision of feedback to data providers

  9. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Other benefits of data monitoring • Conclusion

  10. Monitoring tools • Monitoring tools consist of the routine production and display of statistics over large data samples • Statistics are generally computed for observation quantities related to the data assimilation: departures, bias correction, data counts, etc. • Statistics are produced for various data selection criteria

  11. Monitoring tools • Monitoring tools are designed to produce plots allowing the investigation of data from various perspectives: time, area, vertical, FOV, etc. • Monitoring tools allows generic comparison of statistics from different experiments

  12. Monitoring tools Time evolution of statistics over predefined areas/surfaces/flags

  13. Time evolution of statistics of zonalmeans

  14. Assessment of the geographical variability of statistics: • location effect • air mass effect

  15. FG departure STDV for SMOS YY polarisation at 40 deg J.M. Sabater

  16. compact product for high spectral resolution sounders

  17. Vertical versus latitudes

  18. Time versus longitudes

  19. Solar time statistics

  20. Useful way to compare observed values against model ones

  21. Vertical statistics of time and area averages

  22. Vertical statistics of time and area averages

  23. Vertical statistics of time and area averages

  24. Histograms of time and area averages

  25. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Other benefits of data monitoring • Conclusion

  26. Feedback info (ODB) Current Statistics Per Data type, channel Past Statistics Per Data type, channel Set and adjusted manually Hard limits Detect slow drifts Soft limits Detect sudden changes Anomaly detection Various observation quantities Ignore facility Warning message Web E-mail

  27. Automatic Alarm system Statistics to be checked depend on data type. For instance for radiances: • Numbers • Mean (bias corrected) FG departure • Standard deviation of FG departure • Bias correction • Any other useful observation quantity

  28. Automatic Alarm system Web publishing: • Public access • Published by data types • Time series provided • Severity highlighted • Time limited archive E-mail dissemination: • Subscription by data type • Subscription by severity level • Time series provided

  29. Checking 0001 DCDA 2009012700 ================================= atovs ================================= NOAA-16 AMSU-A 9 clear radiances : out of range: (2 times in last 10 days for at least one item) http://intra.ecmwf.int/users/str/sat_check/atovs_207_3_9_210.png Considerably: stdev(fg_depar)=0.32668, expected range: 0.17(H) 0.29 NOAA-16 AMSU-A 10 clear radiances : out of range: (11 times in last 10 days for at least one item) http://intra.ecmwf.int/users/str/sat_check/atovs_207_3_10_210.png Slightly: avg(fg_depar)=0.019286, expected range: 0.021 0.107 Severely: stdev(fg_depar)=0.427388, expected range: 0.19(H) 0.28(H) NOAA-16 AMSU-A 12 clear radiances : out of range: (5 times in last 10 days for at least one item) http://intra.ecmwf.int/users/str/sat_check/atovs_207_3_12_210.png Severely: stdev(fg_depar)=0.584939, expected range: 0.32(H) 0.48(H)

  30. 20/05/2013

  31. 2013/05/20 12 DCDA 2013/05/21 00 DCDA

  32. Outline • Introduction • Why do we need monitoring? • Monitoring tools • Automatic alarm system • Otherbenefits of data monitoring • Conclusion

  33. Otherbenefits of data monitoring • Data monitoring is mainly based on statistics of FG departures • FG departures are the combination of observation errors, FG errors and observation operator errors. • In most cases, change in statistics is due to change in the characteristics of observations

  34. Otherbenefits of data monitoring • In many cases, change in statistics is due to model problems. • With many satellite providing the same data, it’s possible (via a consistency check) to detect model problems Stratosphere: Microwave, Infrared and GPS data from various satellites. Troposphere: Microwave and Infrared radiances from various satellite Surface: Microwave and scaterrometer data from various satellites.

  35. Sudden stratospheric warming

  36. Early January 2013, the automatic alarm system generated severe alarms associated to an increase of the noise for infrared and microwave stratospheric peaking channels Checking 0001 DCDA 2013010212 ================================= AQUA AIRS 56 radiances : out of range: (3 times in last 10 days for at least one item) http://www.ecmwf.int/products/forecasts/satellite_check//do/get/satcheck/3215/110485?showfile=true Severely: stdev(fg_depar)=0.777, expected range: 0.57 0.68 Slightly: avg(biascorr)=-0.02000005, expected range: -0.37 -0.05(H) METOP-A IASI 89 radiances : out of range: (6 times in last 10 days for at least one item) http://www.ecmwf.int/products/forecasts/satellite_check//do/get/satcheck/3217/111259?showfile=true Severely: stdev(fg_depar)=0.459, expected range: 0.33 0.41

  37. The increase of the noise is due to the onset of the polar vortex breaking out process (SSW). The model predicted the onset of the process but not accurately in the beginning. A lot of data failed the QC check and delayed the model recovery. The SSW was a clear indication that a cold spell will hit Europe one to two weeks later.

  38. Sea-ice monitoring

More Related