This recipe extends the available tutorial documentation on LAS (see https://podaac-tools.jpl.nasa.gov/drive ... orials/las, and illustrates the usage of the Live Access Server (LAS) web-based visualization and subsetting tool to extract a time series of mean values for an area and time period of interest. In this example, we focus on a L4 blended sea surface temperature (AMSR-OI SST) dataset (see https://podaac.jpl.nasa.gov/dataset/NCDC-L4LRblend-GLOB-AVHRR_AMSR_OI) and a time series of statistical averages for the region of the Aegean within the Mediterranean Sea.
In general and were one to do this outside of LAS, several steps would be involved:
1. PO.DAAC maintains numerous SST datasets (see https://podaac.jpl.nasa.gov/datasetlist?ids=Measurement&values=Ocean%20Temperature), many of which could be processed to give one the desired time series of SST values. So the first step is the selection of the appropriate data product(s) given criteria such as sensor type, spatio/temporal resolution and extent relative to the needs of the specific science problem you are trying to address. Note this dataset selection and understanding and stage applies also even if one is using a tool like LAS to undertake the mechanics of the dataset processing and extraction.
2. Downloading of the data files via FTP or using an available web-service such as OPeNDAP.
3. Reading the data files into computer memory, the majority of which are global coverage data, and then spatially subsetting the data for your region of interest (OPeNDAP allows this spatial and temporal subsetting to happen as part of the URL request; thus you only download data spanning the area and time period of interest rather than needing to acquire a whole series of individual global files that then need to be read and subset by you).
4. Once the subsetted data for your area and time period of interest are read into computer memory in some kind of analysis package (eg. MATLAB, IDL), you would then need to do some computations on the data, most likely by running an automated script, to give you the time series of descriptive statistics (eg. mean, variance etc) that you are after.
As one can see this is a rather involved, and may be overwhelming for someone new to working with satellite datasets. Fortunately, however, LAS does much of the above mechanics for you and requires no programming expertise or specialized software other than your browser. LAS can be accessed at ...
STEP 1: Dataset/parameter selection
1. Click the "Choose Dataset" button on the LAS main menu, and then click"+" to expand "Ocean Temperature" and see available datasets
2. Click and expand the “GHRSST” Version dataset grouping and then select "AMSR_OI" from the list to select it (list item #1).
3. Then click "analyzed SST" from the list of available data variables in the AMSR_OI dataset for display.
STEP 2: space/time filter and time-series analysis specification
1. Specify the spatial bounding box for your region of interest by either typing in precise N-S-E-W coordinates in the coordinate input fields (left panel of LAS), or use the interactive mini-map tool to specify the general location of your area using your mouse to draw the bounding rectangle. For the Aegean I specified the following coordinate range: N41.4 -> N34.4, E21.71 ->E28.65 (these are approximate but for your purposes should suffice.
2. Specify the plot type as Time Series (left panel, underneath the mini-map) - the default setting is to view spatial data maps, but under LinePlots, you can click Time Series to get XY output of variable versus time.
3. Next specify the time range of interest in the Date Selector fields at the bottom (left panel), which in your case will be the full time series available for this AMSR_OI dataset. (Note that although this is one of the longer-running datasets available via LAS, starting in 2002 and ends in 2011. If you need to extend the time period then you will have to explore the other SST data available on LAS and extract additional time series as we are doing here but for the more recent period in the other dataset. Then you can try to match both series by normalizing to a common mean SST baseline).
4. Now specify the time series "Analysis" options further below: AnalysisType=Average, AnalysisRegionType = Area (this computes a time series of average values for your specified region on the time interval of the dataset (ie. Daily for AMSR)
5. Finally click the "Update Plot" in the main menu to update the main map window to start processing the data. (Note: check the Update Plot check-box if you want the plots to automatically update when new selections are made - advisable for maps.)
6. LAS will then start processing your data, something that can take some time. There is thus an option, as shown in the figure below, for you to specify your email address to receive and automated notification once the job is done. (Note that sometimes the automated notification emails land up in one's JunkEmaill/Spam folder, so check that if nothing appears to arrive after a while). If you select the notification option then eventually you will receive an email that will contain an extended URL for the AMSR over the Aegean that you can copy and paste in your Browser to see the Time series plotted in.
7. Alternatively, if you don’t ask for notification, then the time series plot will appear in the LAS browser window once the query is executed as shown here:
8. In both cases, it is not only the plot graphic that is available but the actual underlying computed data series that can as a next step be exported for download to a file in a variety of formats (eg. ASCII/CSV). To export the computed AMSR_OI Aegean mean SST Time Series data click the "SAVE AS" button on the main menu (top). The figure below shows the pop dialog box used to then specify export data format and time ranges (select ASCII/CSV, and for the latter accept the default/maximum time range). The resulting output LAS text file for the example query described can be reviewed here:
Finally, as mentioned above, you can repeat this sequence of steps for other SST datasets available via LAS to try extend the series further. However, you may then need to normalize both to a common level based on any overlapping time period since SST baseline values may shift slightly for different sensors. Or you could more simply present both of these as different series in parallel.