This project has moved. For the latest updates, please go here.

ODM design for Variables, Time support

Sep 6, 2011 at 9:32 PM

In the ODM, a new variable must be defined if you change time support for your instrument.

We changed from 15 minute collection to hourly collection. this required that I defined a whole new varialbe for Specific Conductivity. This is problematic for outputs, graphs, reports, hydrodesktop, user training. It is actually the same variable, just collecting it less often.

I want to submit an official request to change odm in a future version so that time support is not part of the variable record. We want to graph the 15 minute data with 1 hourly data etc.

You should also consider removing the definition in the variable record of continuous vs. average. I would still want to graph averaged values with continuous values if I had both.  This also should not be part of the variable record.

 

thank you!

Developer
Sep 8, 2011 at 9:00 PM

Kathleen,

I believe that this is part of a bigger question.  I believe that measurement time series and analysis time series are not the same thing.  The current design of ODM has only one type of time series that is used for both purposes, and we have foreced it to fit to both purposes.  As we analyze the data needs of the community as part of revising ODM, I believe that some way of managing measurement metadata such as time support is needed seperately from the analysis groupings.  Kathleen's issue is one example of this.  Another example is evident with the vertical offsets.  Here we sometimes have the opposite problem, where we would prefer to analyze data with different offsets seperately, even though the data points use the same variable.  So, this is one of several issues that are being considered as part of the next ODM design revision.

Kim

Developer
Sep 8, 2011 at 10:32 PM
Translation: Data management needs are different that data analysis needs. Internal data needs are different that external datasets (raw vs processed and quality controlled).
Kat's thoughts better match the management of data done by Kisters and the USGS. Observations are associated with a sensor at a location, and different processing creates additional streams/series. In the USGS, the same "variable" say 00060 means the same thing across all datasoruces: real time, Daily, and sproadic. In fact, the real time sampling rate of a station may change if they are trying to capture an event.
kimschreuders wrote:

Kathleen,

I believe that this is part of a bigger question.  I believe that measurement time series and analysis time series are not the same thing.  The current design of ODM has only one type of time series that is used for both purposes, and we have foreced it to fit to both purposes.  As we analyze the data needs of the community as part of revising ODM, I believe that some way of managing measurement metadata such as time support is needed seperately from the analysis groupings.  Kathleen's issue is one example of this.  Another example is evident with the vertical offsets.  Here we sometimes have the opposite problem, where we would prefer to analyze data with different offsets seperately, even though the data points use the same variable.  So, this is one of several issues that are being considered as part of the next ODM design revision.

Kim

 

Sep 23, 2011 at 9:48 PM

This is an important discussion. Some of the tools use the data series stored in the database that DavidV refers to as designed for internal data needs. But the users are having to choose betweem and decipher these multiple dataseries and it is a barrier for data analysis and tool usage. ODM Tools and HydroExcel use these data stored in multiple variables with different time support, e.g., and users must know to look for multiple variable codes and data series for what are the actually the same variables (as the scientific community defines the word variable).  I am happy this is being considered as part of the next ODM revision.

Sep 26, 2011 at 3:07 PM

I have a database

SFE_CTDsondes to track a network of sondes we have in a basin. Some sondes are in groundwater, some are in surfacewater.  Some have QC = 0, some have QC = 1. Some data points are 15 minute, some are hourly. I have to create multiple variables for specific conductivity. Then when I use ODM tools or Hydroexcel, these datasets are all separated and not viewable togehter on a graph without doing programming:

GWSpec_Cond_Hourly

GWSpec_Cond_15

SWSpec_Cond_Hourly

SWSpec_Cond_15

As a user in Hydrodesktop for example,  I would like to choose Sites that are groundwater sites, then see all the Specific Conductivity. The variables table should not have to worry about wether the info is collected in GW or sW, 15 min or 30 min.

Maybe Sample Medium should be associated with a site record? is a groundwater site, or a surfacewater site?