Basic Data Manipulation

Using the Data Catalog

Selecting the dataset

The first step is to choose a dataset. The opening page offers several ways to search for a dataset: by keyword search, by category, or by (local) name. Several categories are given; within each category is a list of datasets. For example, one category is the Air-Sea Interface, which contains oceanic and atmospheric data at the ocean surface. There, for example, you will find a link to Surface Climatologies.

Surface Climatologies is a list of datasets that have been averaged to represent a normal year at the air-sea interface. One such dataset is the OSU Surface Data.

Using the dataset

Having selected a dataset, you now get a page that describes the data you have selected, and gives you a number of options for processing that data. Among other things, you can quickly get copies of the data, either as netCDF or HDF files; you can scan through the data with the Data Viewer; you can choose from a number of different possible plots; or you can select a subset of the data.

The Data Viewer allows you to select the data you want by clicking on plots of the data, zoom in and out, control plot size and aspect, etc; it requires a WWW client with in-page graphics and forms (see list of viewers). It has a help page that explains some of its many features.

All the other pages are fully functional for almost all WWW clients. In this case (OSU surface data), if you cannot use the data viewer it makes sense to select a subset of the data before plotting. Otherwise, we end up with many too many plots.

Example: March SST

For example, suppose we wanted to look at March Sea Surface Temperature from the OSU data set. We first select the data set (here). We then select the data we want by choosing

Data Selection, sst, T, Mar, VALUES, Stop Selecting.

At this point we are back to a dataset description page that at first glance looks much like the page we got when we first picked the dataset. A closer look reveals that the page is different: only the March SST is available. We could now get a netCDF file by choosing netCDF, etc, anything that was available on the full dataset is available on this subset of the dataset.

We now select the plot we want by choosing
Plots, .gif, and (under 'Color plots') X Y.
Alternatively, we could choose under 'Contour plots' to get contours.

Example: March SST rearranged

At this point, you might think that while the plot is very nice, you would rather have the plot edges go through land instead of the middle of the Pacific. So back up out of the plot page (here), and choose

Data Selection, X, 20E, 25E, 15E, RANGE, Stop Selecting.

If we now make the same plot selection ('Plots, .gif, etc'), we get the color or contour plot.

Example: March SST in the Pacific or North Atlantic

Similarly, we could pick out a range of longitude (X) and latitude (Y) that gives the March SST for just the Pacific or just the North Atlantic.

Example: Yearly averaged SST

Instead of using the March SST, we could look at the yearly average SST. Going back to the original dataset, we choose

sst, Filters and 'Find the average over T'.

If we now fix the land split and make the same plot selection ('Plots, .gif, etc') as before, we get OSU which we can compare to the Levitus surface climatology with the same land split, Levitus. Clearly the Levitus data has much more resolution, but the two datasets essentially agree.