Cloud Top Properties
Background
This section describes a set of four algorithms that determine cloud top properties. This includes the following algorithms:
Cloud Top Temperature, Height, and Pressure
The Cloud Top Height algorithm will use ABI infrared bands to simultaneously retrieve Cloud Top Height, Cloud Top Temperature, and Cloud Top Pressure for each cloudy pixel. These cloud products are a prerequisite for generating other downstream products that include the Cloud Layer product, Cloud Optical/Microphysical products, and the Derived Motion Wind products. Forecasters will be able to use this information to determine areas of cloud growth and likelihood of precipitation. Other operational applications of this product include its use in Aviation Terminal Aerodrome Forecasts (TAFs), supplementing upper-level cloud information to the ground-based Automated Surface Observing System (ASOS), and initialization of clouds in numerical weather prediction models.
Cloud Top Phase
The Cloud Type algorithm will use four GOES-R ABI infrared spectral bands to determine four different cloud phases: warm (>0C) liquid water, supercooled liquid water, mixed, and ice. The Cloud Phase product is a prerequisite for generating other downstream products that include Cloud Height, Cloud Optical Properties, Fog Detection/Depth, and Aircraft Icing. The Cloud Top Phase product will enable meteorologists to better monitor and track changes in the water properties of clouds, improve icing forecasts for the aviation community, and aid in improving warnings for severe weather. Cloud Phase product information can also be used in advanced ABI applications such as severe weather prediction and tropical cyclone intensity estimation.
Cloud Top Optical and Microphysical Properties
The Cloud Effective Particle Size will be computed using the same algorithm that estimates the Cloud Optical Depth. Using both the visible and near-infrared bands during the day and the infrared bands during the night, the GOES-R Cloud Optical and Microphysical Properties algorithm will retrieve, simultaneously with COD, the Cloud Particle Size. The Cloud Particle Size will provide valuable information about the radiative properties of clouds. This information combined with the information provided by the COD product will provide very accurate information about the Earth's radiation budget, yielding more accurate climate prediction possibilities.
Product Description
Cloud Top Temperature, Height and Pressure
The ACHA is responsible for estimation of vertical extent for all cloudy ABI pixels. In terms of the F&PS, it is responsible directly for the Cloud-Top Pressure, Height and Temperature products. The cloud height is also used to generate a cloud-layer flag which classifies a cloud as being a high, middle or low-level cloud. This flag is used in generating the cloud-cover layers product.
Cloud Top Phase
The cloud type product consists of 6 cloud classifications. In addition, the cloud phase product consists of 4 cloud classifications. The cloud type categories are: warm liquid water cloud, supercooled liquid water, mixed phase, opaque ice, cirrus (e.g. semi-transparent ice clouds), and multilayered cloud (with semi-transparent upper-layer). The cloud phase categories are: warm liquid water phase, supercooled liquid water phase, mixed phase, and ice phase. The cloud phase is directly derived from the cloud type categories. The cloud type product contains information on multilayered clouds and cirrus that is useful to higher-level algorithms such as the cloud top height retrieval.
Cloud Optical and Microphysical Properties
GOES-ABI daytime microphysical properties (DCOMP) have the retrieved properties of Cloud Optical Depth (COD), Cloud effective Particle size (CPS), Liquid and Ice water path (LWP and IWP). Daytime pixels are defined as all observations with a solar angle of 65 degrees or lower.
The GOES-R nighttime microphysical properties (NCOMP) algorithm is responsible for the calculation of water/ice COD, CPS and water/ice path for all ABI nighttime cloudy pixels. In our context, the determination of nighttime is defined to be where the solar zenith angle for a given pixel is greater than or equal to 90°. In addition, these same cloud properties are calculated for solar zenith angles greater than or equal to 82 degrees and less than 90 degrees, but only in a qualitative sense. Another point to keep in mind is that the current algorithm design utilizes cloud phase (inferred from ABI cloud type) and cloud top temperature. Cloud types and cloud top temperatures are determined by ABI algorithms that must be invoked prior to running the algorithm. An attempt will be made to derive COD, CPS, LWP and/or IWP for all pixels that are cloudy with quality flags indicating the degree of success.
How does it work? - Algorithm
Cloud Top Temperature, Height and Pressure
The cloud top properties algorithms uses the infrared observations from the ABI to extract the desired information on cloud height. Infrared observations are impacted not only by the height of the cloud, but also its emissivity and how the emissivity varies with wavelength (a behavior that is tied to cloud microphysics). In addition, the emissions from the surface and the atmosphere can also be major contributors to the observed signal. Lastly, clouds often exhibit complex vertical structures that violate the assumptions of the single layer plane parallel models (leading to erroneous retrievals). The job of the ACHA is to exploit as much of the information provided by the ABI as possible with appropriate, computationally efficient and accurate methods to derive the various cloud height products.
The formal Algorithm Theoretical Basis Document, (PDF, 7.95 MB) is available
![Example of the Cloud Phase product](images/cloud_images/baseline-cloud-phase.jpg)
Example of the Cloud Phase product as generated by the GOES-R Cloud Phase algorithm using Meteosat-8/SEVIRI data on 10 August 2006 12:00 UTC.
Cloud Top Phase
The ABI cloud type/phase algorithm utilizes a series of spectral and spatial tests to determine the cloud type (liquid water, supercooled water, mixed phase, optically thin ice, optically thick ice, and multilayered ice). The algorithm utilizes ABI channels 10 (7.4 μm), 11 (8.5 μm), 14 (11 μm), and 15 (12 μm), which are all infrared channels. In lieu of brightness temperature differences, effective absorption optical depth ratios are used in the spectral tests. Effective absorption optical depth ratios, allow for improved sensitivity to cloud microphysics, especially for optically thin clouds. The validation analysis indicates that the algorithm with comfortably meet the accuracy requirements.
The formal Algorithm Theoretical Basis Document, (PDF, 8.07 MB) is available
Cloud Optical and Microphysical Properties
There is one major product produced by the GLM software: a lightning dataset. To obtain this dataset, the satellite data stream needs to be decoded, filtered, clustered, and output to the appropriate file. The LCFA only generates the lightning dataset. Specifically, the LCFA receives as input the Level 1b pixel-level optical "event" data and processes this data into more convenient lightning data products that are easily utilized by the scientific research and broader operational user communities. Therefore, the LCFA must take the event data and assemble the higher level clustered lightning data products (events, groups and flashes), and in so doing, it will generate derived lightning characteristics associated with these higher level products. It will also interrogate individual flashes, groups, and events on a statistical basis to see if they are associated with lightning or noise [i.e., the Lightning AWG Team does not assume that the noise filtering performed by the Instrument Vendor is perfect]. Definitions of the basic data storage classes (events, groups, flashes) that drive the LCFA are provided below.
The formal DCOMP Algorithm Theoretical Basis Document, (PDF, 13.22 MB) is available
The formal NCOMP Algorithm Theoretical Basis Document, (PDF, 19.54 MB) is available
How are the results compared to existing data? - Calibration and Validation
Cloud Top Temperature, Height, and Pressure
CALIPSO cloud-top height and MODIS CO2 slicing results are used for ACHA validation. In the GOES-R era, the cloud top from the LIDAR onboard EarthCare, along with the JPSS (VIIRS and CrIS) cloud height products will be used.
A more technical validation presentation, (PDF, 24.04 MB) is also available
Cloud Phase
The CALIPSO cloud phase product has been determined to be valid to use for validation of the ABI Cloud Phase product. However, in the GOES-R era, when new space-based LIDARs become available, the LIDAR cloud boundaries along with NWP temperature profiles will be used to identify clouds which have a cloud top temperature less than 233 K (definite ice clouds) and clouds which have a cloud top temperature greater than 273 K (definite water clouds). CloudSat, or future space-based cloud radars, are other potential sources for validation.
A more technical validation presentation, (PDF, 24.04 MB) is also available
Cloud Top Optical and Microphysical Properties
Daytime Validation
Validation consists of comparison to same products from other algorithms on similar sensors and various datasets from field campaigns. This includes applying the ABI DCOMP algorithm to other (similar) sensors, such as SEVIRI, MTG and MODIS, and comparing the results to algorithms developed by other groups (example: EUMETSAT Cloud Retrieval Evaluation Workshop). In addition, the liquid and ice water path (which can be derived from the cloud optical depth and effective radius), can be validated against data from microwave sensors such as AMSR-E and AMSU over ocean surfaces. Cloud transmission (an intermediate product) can be compared to surface flux measurements from the SURFRAD network. Validation of LWP also offers performance estimates of cloud particle size since LWP is generated using both optical depth and particle size. Independent measures of particle size are lacking.
Nighttime Validation
Microwave imagers also provide LWP (derived from the optical depth and effective radius) data for comparison over the ocean. Microwave sensors are sensitive to most values of IWP (derived from the optical depth and effective radius). CALIOP and other space based LIDARs will provide cloud optical depth and IWP for thin cirrus. CloudSat also provides IWP values that are useful for some NCOMP retrievals.
A more technical NCOMP validation presentation, (PDF, 10.5 MB) is also available