tmp Documentation - GerardoLopez/TATSSI GitHub Wiki

2.1 Translator

TATSSI allows converting to/from all GDAL supported formats. This is mainly an internal module that is used to translate input files into an Cloud Optimised GeoTIFF COG which is the default internal TATSSI format. It is possible to use the Translator module from a Python script, an example is provided on the TATSSI (Translator)[https://github.com/GerardoLopez/TATSSI/blob/master/TATSSI/notebooks/TATSSI_ImportExport.ipynb] Jupyter Notebook.

An example of how to use the module to convert the MOD13A2 EVI SubDataset product into a COG is shown below:

# Import directly the TATSSI translate module
import os
from glob import glob
from TATSSI.input_output.translate import Translate
from TATSSI.input_output.utils import *

# Set data directory
DataDir = "/home/TATSSI/data/MOD13A2.006/"
# Get all HDF files into a list
fnames = os.path.join(DataDir, '*.hdf')
fnames = glob(fnames)
# Sort the list
fnames.sort()

for fname in fnames:
    # Get EVI QA SubDataset (SD) which is index 1
    sds = get_subdatasets(fname)
    vi_qa_sds = sds[1][0]
    
    # Set output file
    directory_name = os.path.dirname(os.path.abspath(fname))
    output_fname = os.path.join(directory_name,
                       os.path.basename(fname)[:-3] + 'EVI.tif')
    
    # Extract to a GeoTiff file
    Translate(vi_qa_sds, output_fname, 'GTiff')

3.1 Time Series

All files available at the LP DAAC are available on a file per time step, either daily, 8-day, 16-day or monthly data. This is not the most efficient way to have the data when analyse time series where per-pixel operations for every time-step are required. Additionally, in order to process time series TATSSI requires specific metadata to set the observation/acquisition time for every time set in the time series.

Several of the MODIS and VIIRS products distributed on the LP DAAC are in HDF4 or HDF5 formats, given the hierarchical nature of these formats, this means that a single file can contain multiple subdatasets, we will referred to them as Scientific Datasets (SDS).

TATSSI will create a time series for a product where the default behaviour is:

For every file in the data directory that matches the product selected by the user:

Will import each band or SDS to the internal TATSSI format - Cloud Optimized GeoTiff (COG)

For each QA layer associated to the product:

Will import it to the internal TATSSI format and perform the QA decoding

For each band or SDS and associated QA layers

Will create GDAL VRTs layer stacks

3.2 Select spatial subset

TATSSI can either generate the time series for the full geographic extent or a spatial subset of the product selected. To generate a spatial subset:

Click on the Select spatial subset button to show the TATSSI time series generator - Spatial Subset window as shown in Figure 3.2.image1
Click on the Zoom to rectangle button, the cursor is going to change to a cross, to select the subset, click and drag to draw a rectangle, the map will now show that spatial subset as shown in figure 3.2.image2. Click on the Get extent for subsetting to indicate that will be the subset to use and then close the window.

3.1.3 Format | Driver | Extension

TATSSI selects the Format and GDAL Raster Drivers using the extension of the first file in the data directory selected by the user. It could be the case that the files do not have a consistent file name extension and corresponding format, if that is the case, select the adequate Format | Driver | Extension.

3.2.6 Graphic controller buttons

Save the figure Saves the current plot into a local image, e.g. PNG or JPEG file.

3.3.7 Standard anomalies

A_t = (X_t - mu_t) / std_t

https://github.com/GerardoLopez/TATSSI/blob/e2abafc14d2c1c644a60aa956229175711da371d/TATSSI/UI/time_series_analysis.py#L589

4.1.10 Save QA Analytics

Saves the current selection for the selected QA definition and corresponding QA parameter name into a JSON file. This file can be then used to apply the same QA settings to a different time series.

4.1.13

Shows the histograms for the % of data available and the ``Max-gap length```