Single footprint satellite product problem in CERES

[xwgw]

Help post: Single footprint satellite product problem in CERES
What is the cloud size of SSF in the single footprint satellite product of CERES, cloud_fraction (PS: there is a PSF-wtd MOD04 cloud fraction land in its product)? Because there are two cloud products in the grid product of the single footprint satellite product (PS: Cloud Area Fraction and MODIS Land Aerosols Cloud Fraction), I would like to obtain information about the cloud cover size of different clouds, which I should get. I hope to get your help. Thank you very much.

[wfmiller]

The MODIS Land Aerosols Cloud Fraction is there to assist the user and interpreting the MODIS aerosol optical thickness. Since the MODIS aerosol team uses a different cloud mask than the CERES project. They use a 10 by 10 MODIS pixel grouping and determine the cloud fraction within it. There are multiple 10 by 10 MODIS pixel groupings within a footprint, so we provide the average value over a footprint. If you are using the gridded hourly product, than the footprints values within a grid box are averaged. If the land aerosol cloud fraction is large, than there is a high probability that the aerosol optical thickness have some cloud contamination.

The cloud size is not provided in the SSF product since it may be larger than a footprint or contain multiple small clouds within it. If you are using the gridded product and assume uniform sampling of the grid box, then the size (area) if the cloud would be the area of the grid box (1 degree latitude by 1 degree longitude) multiplied by the cloud fraction. This assumption is not bad for synoptic and uniform clouds like a stratus region. It would hold up less in cumulus fields where the individual clouds are much smaller than the total area they would cover in the grid.

If you are using individual SSF footprints, you would need to use the view zenith angle and satellite radius to determine how large the footprint is before using the cloud fraction. Again, the cloud may not be contiguous even within a footprint.

[xwgw]

Hello, thank you for your help. But I'm a little confused, because I think we need to estimate the surface radiation budget. But it is much more complex than estimating TOA because it requires a radiative transport model and satellite-derived cloud and aerosol properties as well as atmospheric states from satellite or reanalysis. The basic assumptions and the errors of the auxiliary input data in the radiative transfer model increase the uncertainty of the estimation of the surface radiation budget. And I need to input the information of cloud parameters, cloud height value, cloud optical thickness value, cloud droplet particle radius information, some parameter information of aerosol, and atmospheric parameter information into my radiation transfer model (PS: My radiation transfer model is SBDART).
Since I've been using the Footprint satellite Level 2 product data SSF from CERES, Then, cloud parameter data (cloud top height, cloud base pressure, cloud visible optical thickness, cloud liquid water particle radius, cloud ice water particle radius, liquid water path, ice water path) and aerosol parameter data information (MOD04 aerosol types land, MOD04 cloud fraction) are used land, MOD04 corrected optical depth land (0.550), MOD04 deep blue aerosol optical depth land (0.550), MOD04 deep blue angstrom exponent land, MOD04 deep blue single scattering albedo land (0.470); Atmospheric parameter data (surface temperature, surface pressure, altitude, relative humidity and other data information).
But now I have a problem, that is, when I input the cloud parameters into the SBDART model, I do not know how to determine the actual optical thickness of the cloud, and the actual particle radius of the cloud, and the actual height of the cloud. Because there are several channels of data value information for cloud optical thickness, They are Mean visible optical depth for cloud layer, Mean logarithm of visible optical depth for cloud layer, Mean logarithm of visible optical depth for cloud layer (1.2), Mean logarithm of visible optical depth for cloud layer (2.1); For cloud particles the radius is divided into ice water particles and liquid water particles, They are Mean ice particle effective radius for cloud layer (1.2) and Mean ice particle effective radius for cloud layer (2.1), Mean ice particle effective radius for cloud layer (3.7), Mean water particle radius for cloud layer (1.2), Mean water particle radius for cloud layer (2.1), Mean water particle radius for cloud layer (3.7), Mean ice water path for cloud layer (3.7), Mean liquid water path for cloud layer (3.7), mean ice water path for Cloud Layer (3.7), For the cloud fraction of the cloud amount only MOD04 cloud fraction land is provided.
I want to know how I can choose to get the cloud optical thickness and cloud particle radius information? (Note: Because I saw in the SSF2 data quality summary that the optical thickness of the cloud is generally measured using a visible light channel of 0.65 microns, and the cloud particle size is expressed using a 3.7 micron channel), when I want to get the optical thickness of the cloud, Can you directly select Mean visible optical depth for cloud layer? To obtain cloud particle radius information, Can you directly select Mean ice particle effective radius for cloud layer (3.7), Mean water particle radius for cloud layer (3.7)? 3.7 micron channel ice particle radius and water particle radius? Since the cloud parameter and aerosol parameter information provided by modis are matched with the CERES field of view, can this MOD04 cloud fraction land represent the cloud fraction of clouds under the same field of view?
In addition, through the information of the radius of the ice water particles, can the particle radius information value of the actual cloud directly be obtained? (My idea is to determine whether it is an ice cloud or a water cloud based on the phase state of the cloud. If the ice cloud is directly used, Mean ice particle effective radius for cloud layer (3.7). Mean water particle radius for cloud layer (3.7) Mean water particle radius for cloud layer (3.7)?
Sorry to bother you, my question is very trivial, but I really need to understand these situations, I hope you can help me, thank you sincerely, and look forward to receiving your reply as soon as possible.

[wfmiller]

Yes, I understand the large amount of data need for running a radiative transfer model. The CERES project runs the Fu-Liou RTM for several products, CRS, CRS1deg-Hour, SYN1deg-1Hour, SYN1deg-MHour, SYN1deg-Day, and SYN1deg-Month. The CRS is a SSF based product where the others are gridded products.

How much of the cloud data on the SSF you used is tied to how complex the radiation transfer model is and the resulting accuracy. Hence these decision are left to the user.

There is a variable “Clear/layer/overlap percent coverages” which has “Conditions:clear, lower, upper, upper over lower.” The second and third variable is the cloud fraction of the lower and upper layer corresponding to the other cloud properties on the SSF. Again, the MOD04 cloud fraction land is something carried over from MODIS aerosol product and not tied to the CERES cloud mask and properties provided on the SSF. The Mean cloud particle phase for cloud layer (3.7) will indicate the proportion of the cloud that is water (1) or ice (2), i.e. near one is water and 2 is ice. The phase will allow you to determine if you want to use water radius and water path or ice radius and ice water path in the model. The Cloud visible optical thickness, cloud droplet water particle radius and ice water particle radius at 3.7, and liquid water path and ice water path at 3.7 microns are a consistent set of measurement. One phase is generally used per cloud in the RTM instead of trying to blend them.

The answer on which set of properties to use will depend on what the RTM you are using is expecting which I don’t know. The variables on the SSF may work fine. If it needs the OD at 3.7, then it can be computed from the particle radius and water path variable that are on the product.

For the RTM used in CERES products, the scientist provided the following:

In the Fu-Liou code, Liquid water path (LWP) is derived using the input effective radius and visible optical depth.

Then the LWP is converted into the cloud optical depths at the Fu-Liou bands using the scattering parameter database (the ratio between the cloud extinction coefficient to cloud water content).

A similar process is also applied to the ice phase – first ice water path (IWP) is retrieved using ice particle effective radius and visible optical depth.

Then IWP is converted into the optical depth at the Fu-Liou bands.

[xwgw]

Hello, sorry to bother you again. I have another question to ask you, that is, for the short wave, long wave and surface and top atmosphere radiation flux information provided by CERES. For short wave, I know that the wavelength range of short wave is 0.2-5 microns, and the window channel is 8-12 microns. Long wave channel I look at that product description (CERES_SSF_Terra-Aqua_Edition4A number
According to the quality abstract, the long-wave channel radiation flux said in the daytime is the total wavelength channel 0.2-100 microns minus the short-wave channel 0.2-5 microns, which is not explained at night, because I want to take a separate radiation transmission mode to simulate the short-wave and long-wave radiation flux information of the surface and the top of the atmosphere, I want to understand in detail. How is the long and short wave radiation flux information provided by CERES obtained? What is the wavelength range of the long-wave flux?
Look forward to your help, thank you!

[wfmiller]

The received shortwave energy at night is assumed to be 0. Therefore, the longwave is just from the total channel. Since there should be minimum energy in the 0 to 5 microns area being emitted or reflected by the Earth at night, the day and night values are comparable. The longwave range is considered from 5.0 to 100.0 microns.