Australian VLab Centre of Excellence National Himawari-8 Training Campaign Introduction to the Derived Products Compiled by Bodo Zeschke, BMTC, Australian Bureau of Meteorology, using information from various sources, June 2015 Should you use these resources please acknowledge the Australian VLab Centre of Excellence. In addition, you need to retain acknowledgement in the PowerPoint slides of EUMETSAT, the Japan Meteorological Agency, the Bureau of Meteorology and any other sources of information.

Learning Outcomes At the end of this exercise you will: Have gained a basic knowledge how the Derived Products are constructed from satellite data and Numerical Weather Prediction (NWP) models. Have gained a basic understanding of the "Cloud Type" Derived Product from NWC-SAF Have gained a knowledge of the Derived Products planned for use by the Australian Bureau of Meteorology once Himawari-8 data becomes available. Have gained a basic understanding of the advantages and limitations of Derived Products to the Forecaster.

RGB Image Derived Product (TPW) NWP (TPW) Real “Semi” Real Not Real Derived Products (the Bureau plans to implement Himawari-8 Derived Products from September 2015) All NWP is wrong but some NWP is useful… It may look great but it’s not real… don’t be seduced! from “RGB Products versus Derived Products” Dr. Jochen Kerkmann, presented at WMO EUMETSAT RGB Satellite Products Workshop hour Microphysics RGB

Introducing Derived Products (1) Derived Products are generated: Using selected single band and band combinations of satellite data Applying some manipulation to this satellite data (statistical analysis, applying thresholds etc.) Utilising input from Numerical Weather Prediction models (NWP). Derived Products: Usually depend on some basic assumptions Provide quantitative information and are therefore less subjective than raw satellite data. Take time to compute From RGB Products and Derived Quantitative Products, Marianne Konig and Jochen Kerkmann EUMETSAT

Introducing Derived Products (2) There are different types of Derived Products: Products focussing on cloud-free regions (e.g. SST) Products focussing specifically on clouds (e.g. microphysics) Products using a multitude of information (time sequence, cloud evolution, e.g. Convective Initiation warnings, early warning of storm potential) From RGB Products and Derived Quantitative Products, Marianne Konig and Jochen Kerkmann EUMETSAT

Derived Products we have already looked at (EUMETRAIN ePort) Cloud Type Cloud Top Height Precipitating Clouds Convective Rain Rate

Derived Products we have already looked at (EUMETRAIN ePort) Cloud Type Cloud Top Height Precipitating Clouds Convective Rain Rate

Example of a Derived Product – Cloud Type Product (NWC SAF) ( From RGB Products versus Derived Products J. Kerkmann EumetSAT Fractional and high semitransparent clouds separated using T8.7µm- T10.8µm brightness temperature differences, but also variance in R0.6µm visible reflectance The remaining categories are distinguished through the comparison of their T10.8µm to NWP forecast temperatures at several pressure levels. T7.3µm and T8.7µm are also used to refine the separation between low and medium clouds.

The CT classification algorithm is based on the following approach: Main cloud types are separable within two sets: the fractional and high semitransparent clouds, from the low/medium/high clouds. These two systems are distinguished using spectral features : T10.8µm-T12.0µm, T3.9µm-T10.8µm (in night-time conditions only), R0.6µm (in day-time conditions only), and textural features (variance T10.8µm coupled to variance R0.6µm in daytime conditions). Within the first set, the fractional and high semitransparent are separated mainly using their T8.7µm-T10.8µm brightness temperature differences, but also their R0.6µm visible reflectance (in daytime conditions only). The remaining categories are distinguished through the comparison of their T10.8µm to NWP forecast temperatures at several pressure levels. Example of a Derived Product – Cloud Type Product (part 1) From RGB Products versus Derived Products J. Kerkmann EumetSAT also from

The CT classification algorithm is based on the following approach: T7.3µm and T8.7µm are also used to refine the separation between low and medium clouds, especially useful in case of low level thermal inversion. No separation between cumuliform and stratiform clouds is performed in the current version of CT. A separate processing is applied to compute a cloud phase flag, based on the use of CT cloud type, T8.7µm, T10.8µm (all illumination), R0.6µm and R1.6µm (at daytime only). Example of a Derived Product – Cloud Type Product (part 2) From RGB Products versus Derived Products J. Kerkmann EumetSAT also from

Planned Bureau use of Derived Products (from the Himawari 8 and 9 Project HIM89 Project Definition) PHASE 2A F OG M ASK SUITE [NOAA] V OLCANIC A SH RETRIEVAL SUITE [NOAA] C LOUD P ROPERTIES [NOAA] F OG P ROBABILITY SUITE ( BY LAYER ) [NOAA] V OLCANIC A SH P ROBABILITY SUITE [NOAA] C LOUD M ICROPHYSICAL P ROPERTIES SUITE [NOAA] A TMOSPHERIC M OTION V ECTORS SUITE S OLAR R ADIATION SUITE S EA S URFACE T EMPERATURE SUITE ( PART 1) PHASE 2B A IRCRAFT I CING P OTENTIAL V OLCANIC A SH OBJECT INFORMATION SUITE [NOAA] S EA S URFACE T EMPERATURE SUITE ( PART 2)

Planned Bureau use of Derived Products (in more detail) (from the Himawari 8 and 9 Project HIM89 Project Definition) A TMOSPHERIC M OTION V ECTORS SUITE The AMV suite is being develop by CAWCR and Science and Engineering's Passive Remote Sensing team, and is expected to be finalised before end June AMV Vectors (u/v) NWP Selection Criteria (EE, QI) Data Processing Flags F OG M ASK SUITE [NOAA] Fog Mask Fog Depth F OG P ROBABILITY SUITE ( BY LAYER ) [NOAA] MVFR Fog Probability LIFR Fog Probability IFR Fog Probability IFR RH only Fog Probability V OLCANIC A SH RETRIEVAL SUITE [NOAA] Ash Top Temperature Ash Top Pressure Ash Top Height Ash Emissivity Ash Beta Ash Optical Depth IR Ash Mass Loading Ash Effective Radius S OLAR R ADIATION SUITE Solar Global Horizontal Irradiance Solar Direct Normal Irradiance Solar hourly Global Horizontal Exposure Solar hourly Direct Normal Exposure Solar Daily Global Horizontal Exposure V OLCANIC A SH P ROBABILITY SUITE [NOAA] Ash Probability IR VIS Ash Probability IR Cloud Emissivity Ch14 Cloud Beta 1112 Tot C LOUD P ROPERTIES [NOAA] Cloud Mask Cloud Top Temperature Cloud Top Pressure Cloud Top Height Cloud Type Cloud Phase C LOUD M ICROPHYSICAL P ROPERTIES SUITE [NOAA] Cloud Optical Depth (Visible) Cloud Particle Effective Radius Cloud Liquid Water Path Cloud Ice Water Path Cloud Albedo S EA S URFACE T EMPERATURE SUITE Sea Surface Temperature (Regression Channels 11 and 12) Sea Surface Temperature (Regression Channels 3, 11 and 12) Uncertainty A IRCRAFT I CING P OTENTIAL (PHASE 2) Aircraft Icing Potential S EA S URFACE T EMPERATURE SUITE (PHASE 2) Skin Surface Temperature (Physical Retrieval) [NOAA] V OLCANIC A SH OBJECT INFORMATION SUITE [NOAA] (PHASE 2) Ash Probability Volcanic Convection Detection

RGB Products and Derived Products – central Europe From RGB Products versus Derived Products J. Kerkmann EumetSAT

Met-8, 1 Feb 2007, 01:30 UTC; NOAA-18, 1 Feb 2007, 1:22 UTC 24-h Microphys. RGB Cloud Type (MSG, SMHI) Cloud Type (NOAA, SMHI) Low Clouds / Fog (Night) – different satellite sensors Sweden Slide from “RGB Products versus Derived Products” Dr. Jochen Kerkmann, presented at WMO EUMETSAT RGB Satellite Products Workshop 2012

Cloud Type Products 24-h Micro RGB Low Clouds / Fog (Night) – different algorithms 1 November 2006, 4:00 UTC Arpege Model ECMWF Model Sweden Slide from “RGB Products versus Derived Products” Dr. Jochen Kerkmann, presented at WMO EUMETSAT RGB Satellite Products Workshop 2012

Cloud Type Derived Product – different satellite sensors, different algorithms In the previous two slides you can see some of the differences when using the Cloud Type Derived Product recipe for different satellite sensors and also for different Numerical Weather Prediction (NWP) models. Note that in the case of the different satellite sensors the Derived Product using the NOAA satellite data shows a large area over the Baltic Sea to the east of Sweden as cloud free. The Derived Product using the Meteosat Second Generation (MSG) satellite data shows very low cloud over much of the Baltic Sea instead. In the case of the different NWP models, the Derived Product using the Arpege model shows much of the land surface of Norway. The Derived Product using the ECMWF model indicates that this region is covered by very low cloud and also by some broken cloud. This is an important result because Derived Products can be obtained from both geostationary and polar orbiting satellite data so Forecaster understanding of the strengths and limitations of Derived Products from different sources is important for his/her correct interpretation of this data.

Exercise – compare the Night Microphysics RGB Product and the Cloud Type Product (EUMETRAIN ePort) Question: Give one advantage of the Derived product. Give one disadvantage of the Derived Product.

Advantages of Derived Products over RGB products Disadvantages of Derived Products over RGB products 1 – this data can be calibrated to assist people with colour blindness 2 – less subjective 3 – better for climatology studies 4 – not affected by viewing angle – good at all latitudes. 5 – products can be produced that focus upon particular properties. 1 – loss of texture of the cloud 2 – takes time to compute this – generated later 3 – dependent on NWP and other ancillary information 4 – difficult to animate (often noisy) 5 – not so good for detecting cloud boundaries and thin cloud (thin fog) 6 – reduced horizontal and vertical resolution (thresholding) Question – Advantages and dissadvantages of Derived Products over RGB products

Derived Quantitative Products for Climatology – overshooting stormtop climatology – Eastern Africa (Bedka, NASA) From "RGB Products and Derived Quantitative Products" Marianne König, Jochen Kerkmann,, presented at WMO EUMETSAT RGB Satellite Products Workshop UTC UTC UTC UTC

Summary Construction of Derived Product (satellite data, manipulation of the satellite data, NWP) The Cloud Type Derived Product has been explained in some detail. Derived Products planned for use at the Australian Bureau of Meteorology with Himawari-8 data. The advantages and disadvantages of Derived Products compared to satellite data (RGB products) have been discussed References for further investigation have been given

References EUMETRAIN ePort EUMETSAT NWC SAF WMO-EUMETSAT RGB Satellite Products Workshop _FinalReport.pdf