[Thread Prev][Thread Next][Index]

Re: How to deal with history ocean station data

Title: Re: How to deal with history ocean station data
  I personally would not interpolate these data onto an xyzt grid.  I think you will find that it is much too sparse and interpolation would give users a misleading impression of the actual amount of data that is there.  What we have done with the WOD data is put it onto a grid by averaging it in bins (for sst we use monthly averages, 1-deg squares, and standard depths).  We serve this along with the number of observations in each bin and the standard deviation so users can make an evaluation of the quality of the mean.  If users need data in an area where the data is very sparse, they can then make their own judgement on whether or not interpolation is a scientifically sound thing to do, depending on the nature of their investigation.  If users need data in a box where there is a lot of data - they can see this by the number of observations, they may decide to go back to the actual observations on their own and do further subsetting.  Either way, you are providing them with real data and information rather than a "black box" of interpolated values.   Lynn

At 4:37 PM -0300 6/18/03, Chi wrote:
Hi Joe
      Thanks for your information. Actually my problem is not about the file format of world ocean
database.  I 've got the world ocean view software, used it to import the data I downloaded and export
the nitrate data into a clearly-formatted text file.  All the x/y/z/t  and nitrate values are  well listed in
the exported text file.
      My problem is more of a strategy problem at this stage.  I know the data format, but I don't know
which approach to choose to process the data.  Should I interpolate whole data onto XYZT structure?
I noticed that the numbers of observed data records in different years are dramatically different, for
example, 1700 in 1980 and 70000 in 1995.  Therefore does it make sense to do the interpolation in the
x/y/z/t 4D domain?  Another questions is that the areas where nitrate was measured don't seem to cover
the global ocean.
     In my assumption I'd like to make LAS work on the dataset like this.  The user can select within one of the
areas where data was collected  from the map, also he can choose from the years when data was measured,
then he will get the interpolated data for the area and the year. Therefore I will only need to do interpolation
in an  area in one specific year. Does this idea make sense ?  Besides I don't know how to define multiple
regions in a netcdf file so the user can choose from them in the map?
    Thanks a lot, Ding

Joe McLean wrote:
Hi Ding,

Chi wrote:


Hi dear LAS users
    I have the ocean station data for nitrate from 1965 to 2000, which was downloaded from NOAA world ocean database web site at
    The nitrate data was collected in different areas lon/lat/depth and in different years. For instance, some data was collected in area 1 only at 1970 and 1983. And the areas ( lon/lat) were not uniformly distributed.  I don't know how to proceed on this dataset.

The file you downloaded from NODC contains data in the OCL WOD native format, an ascii representation of the structure and data.  NODC offers some fortran programs which will allow you to read the data.  You can read about this at 


So far my previous LAS experience was that I did 3D interpolation on a profile dataset and managed to make LAS work on it.

In order to interpolate the data onto a XYZT gridded structure, you would need to extract the X,Y,Z,T, and Nitrate values from this file first.

In the LAS developers group, we are working on providing access to the World Ocean Database through LAS and will have more information soon.


    Anybody can give me some suggestions or information? Any help will be appreciated.
    Thanks a lot, Ding

Lynn deWitt


Pacific Fisheries Environmental Laboratories
1352 Lighthouse Avenue 
Pacific Grove, CA 93950-2097

[Thread Prev][Thread Next][Index]

Dept of Commerce / NOAA / OAR / PMEL / TMAP
Contact Us | Privacy Policy | Disclaimer | Accessibility Statement