1

Re: How to deal with replicate samples?

Hi all,

we have quite some surveys that use replicate sampling, for example benthic surveys with up to 15 replica's at one sample location (to cope with spatial clustering of organisms). These replica's are analyzed separately (no bulk samples) and the results need to be stored separately as well. This seems to be a hierarchical level below the sample level that is not represented in the nbn datamodel. Is there a good way to store this kind of data in Recorder? Thank you in advance.

Frederic

2

Re: How to deal with replicate samples?

Frederic

Not sure I fully undertsand, but Sample Relations might help to group together related samples.

Mike

Mike Weideli

3

Re: How to deal with replicate samples?

Hello Mike,

an example:
- a transect with 10 sample points in an intertidal area
- at each sample point 10 soil cores are taken for identification and biomass of soil organisms
- each core is divided in two fractions (250 and 500µm sieves)
- species counts and biomass are extrapolated for each sample point based on the 10 replica's
- data need to be stored for the sample as a whole but also for the individual replica's and fractions

Sample relations are a possibility but there seems to be no option to store the nature (replica, fraction) of the relationship at that level.

Frederic

4

Re: How to deal with replicate samples?

In Recorder terms you need to have one Samples for each soil core/fraction. That is to say in your example you need 20 samples for each sample point (200 in all).  The information about each 'Sample' such as the core number and the fraction are then identified as sample_type, sample_reference  or if necessary by sample measurements.  The issue is then not how to split the samples, but how to group them together. You can do this  with Sample Relation which could be used to link the Recorder samples together to get all the information for one of your samples. Alternatively, you could use the Location Hiercarhy to do this by creating  a Location for your sample point and allocating this to all the core/fraction samples, so that you can bring them together for reporting. Also using the Location hierarchy you could also link together all the samples which make up a transect. 

The best approach is probably to key  examples of the various approaches into the Observation hiercarchy
to see how the data will  look.  I suspect that the Report Wizard is not going to be much use in getting to this data so you will will need to use XML reports.

Mike Weideli

5

Re: How to deal with replicate samples?

Thank you Mike,

using the sample measurements for additional information sounds like a very good solution. Will try this out with a small test dataset.

Frederic