The Time-Series Statistics operation (nicknamed "g3stats") includes the following:
1. Weighted mean. The weighted mean is an unbiased estimate of the mean based on
the following formula:
Where xi are the image observations, wi are the weights depending upon the cosine latitude, and any other type of weighting. The proof that the formula above is an unbiased estimate of the mean is provided on the weighted standard deviation (see PDF document link below). The calculation this estimate was implemented by pygsl the python version of gsl library.
2. Grid Cell Count. The Grid Cell Count calculation provides the number of active grid cells each image contains versus time.
[Grid Cells may be confused with image pixels and the "Pixel Count" variable. A pixel is commonly defined as the smallest data element in remote-sensing data, the spatial area represented by a digital number. Grid Cells may therefore correspond one-to-one with pixels in a data file if the finest spatial resolution is preserved, but in the gridded mapped data commonly used in Giovanni, the spatial resolution is usually reduced by averaging the data over space and time. So the Grid Cell in gridded mapped data is a single data element at the resolution of the mapped data. For example, if the grid cell size is 1 x 1 km and the subset image area is 10 x 10 km, there would be 100 grid cells in this subset image.]
3. Weighted Standard Deviation. The weighted standard deviation is an unbiased
estimate of the standard deviation based on the following formula:
The formula above in was implemented by pygsl which is an alternative version of the gsl library.
4. Maximum. The maximum corresponds to the maximum value of the image. So, if
the image has n observations that would be:
5. Minimum The minimum corresponds to the minimum value of the image. So, if the
image has n observations that would be:
PDF document on g3stats (this includes a proof of the unbiased estimation)
This Technical Summary page is a draft version -- for information on this operation while the documentation is being completed, please send a message to GES DISC Help.