Replies to This Discussion

Permalink Reply by Nicholas John Clarke on March 25, 2015 at 21:45

Although we only use tension lysimeters in the monitoring programme, I have used zero-tension lysimeters in other projects. They were placed in or under the organic layer. At these depths, they could give us samples even in dry periods when tension lysimeters did not work well. Thus, I think there is certainly a place for them in the monitoring programme. Of course, the two lysimeter types sample different water fractions, so samples from zero-tension lysimeters cannot be directly compared with those obtained from tension lysimeters. As long as this is kept in mind, I see no problem.

Nicholas

Permalink Reply by Tiina Maileena Nieminen on March 29, 2015 at 0:26

Thanks Nicholas,

good to hear, I fully agree!

Tiina

Permalink Reply by Elisabeth Graf Pannatier on March 26, 2015 at 9:31

Hi Tiina,

Yes, you have pointed out an important issue that matters for time series analysis.

I think that it is correct to give a new sampler_ID if the chemical analysis corresponds to a new sample (in your case pooled sample). In the SSM file, we have to indicate how many lysimeters are included in the sample and therefore we can see if this is a single sample or a pooled sample. What is missing in the ICP Forests database is the link between the old single sampler_ID (single lysimeter) and the new one (pooled). To make it simple, we could add in the remark field in the PSS file the old sampler_ID corresponding to the new pooled sample.

PS If a lysimeter has to be replaced because it is broken, do we have to change the sampler_ID? I suppose not if it is reinstalled exactly at the same place.

Elisabeth

Permalink Reply by Tiina Maileena Nieminen on March 29, 2015 at 0:24

Thanks Elisabeth,

for your prompt response. Concerning the link between the new pooled and the old sampler-ID, what do you actually mean to add in the remark field. The list of the sampler_IDs which were combined?  We did that now, but I am not sure if it would be of any use. The list is varying in time, because all samplers do not always give water.

For replacement of a broken lysimeter, the same ID can be used according to the current formulation of Manual.

I am starting a week long holidays tomrrow.

Happy easter time!

Tiina

Permalink Reply by Guia Cecchini on March 30, 2015 at 13:22

Dear Tina and dear All,

I’d like to start to think together to the different points you mentioned in your mail.

1)      Pooling the samples.

The drastic decrease of founding  is more or less a real transversal situation, involving  all the working groups.

In Italy we have already experienced a drastic funds reduction and we are pooling soil solution samples since several years.

It’s a problem, as Elisabeth well knows, to manage in order to rearrange a comparable samplers numeration for the data base.

We have to consider  that things are further complicated by changes in the protocol, that have come and gone over the years, without taking into account the needs related to data management.

Now we have data difficult to compare, in a time when there is no money even to work on a possible harmonization. Meanwhile, if we want to keep alive the monitoring program at the European level, we must consider that pooling samples can be the only opportunity to get data from several countries.

So we need strategies that will eventually adopt new ID, easily correlated to those of the existing data base and , as Elisabeth suggests, create a strong link with the PSS remark field , to get like a mini report with detailed explanations. I agree with Elisabeth that, in case of broken lysimeters, we can assume the same ID sampler, if we are going to reinstall more or less in the same place. I’d like to underlain “more or less.  It’s impossible to reinstall exactly in  the same point because to remove the old lysimeter, and just subsitute it, can create problems of contact and tension maintenance. As the water regime is irreversibly compromised  by the broken lysimeter, within about 80-100 cm, we have to reinstall about at that distance. In my opinion it’s a reasonable compromise and we can assume the same sampler ID.

 

2)      Zero Tension Lysimeters.

As Nicholas said, they are used in the organic layers, and, in dry periods, they are the only tool able to collect solutions. Their use is linked to the kind of soil in which we have to work.

Maybe, in some soils with vertic properties, they could be the only chance to get samples, or, for example, they could be easily used in boggy soils, where the different “water fractions” that Nicholas mentioned, is not a problem.

I agree with him. It’s anyway not a problem; it depends on the use, and we have to consider to use all the possible tools for different situations and applications   

 

Happy Easter

Guia

 

Permalink Reply by Nathalie Cools on April 7, 2015 at 22:21

Dear all,

Concerning the pooling of samples, we agreed at the FSEPM in Slovenia that the IDs of the lysimeters, included in the pooled samples have to be included in the "observation field" (so last field of the PSS file).

Though from the user's viewpoint of the database, I start doubting whether this is a good solution. The 'observation field' might contain any comment from any kind and it is not a field which is mandatory to be completed. As the problem of numbering has given (and still gives) many problems, I believe that the reporting of the sampler IDs in the pooled samples should be MANDATORY information to upload for each NFC to the central database. In case this varies in time and can change from sample collection to sample collection, it should be illustrated for each sample collection. To be able to do this, we need a proper reporting format. 

I think it would be worthwhile to discuss with Till Kirchner whether a new, additional file could be uploaded containing this information in a well structured way and easy to access and understand for any user of the data (and link with the other soil solution data files).

What do you think of this idea?

Kind regards,

Nathalie

Permalink Reply by Guia Cecchini on April 8, 2015 at 10:06

Dear Nathalie, dear All,

I think it’s a  good idea. I filled for years the column " observation field  " , and I have always been requested explanations on ID numbers. Probably the program is unable to maintain a number of characters compatible with a detailed explanation on bulks . A mandatory separate report,  with an encoded format strictly linked to PSS and SSM forms, might be easy to manage and to update without having to revise the PSS  and SSM format. We have anyway  to try every possible solution to solve this problem. If we continue to accumulate data, year after year,  in this situation , the database will become unmanageable.

Best regards

Guia

Permalink Reply by Elisabeth Graf Pannatier on April 9, 2015 at 11:16

Dear Nathalie, hi everyone,

I agree with you Nathalie that the observation field is not really useful, especially when working with thousands of data. I also think that we need a field including all Sampler_Numbers used in the pooled samples.

We can do it at two levels of precision:

1. Report the Sampler_Numbers available for the pooled samples in the PSS file. It does not mean that all Sampler_Numbers had water. This is a similar approach as the Number of samplers in the PSS file submitted since 2011. This number is a constant.

2. Report the Sampler_Numbers effectively used for the pooled samples in a separated file (because it varies from sample collection to sample collection)

In Switzerland, we have both information in our database. Frankly speaking, I use mainly the first approach. Using the second approach is very difficult because I never know whether the change is due to the lack of water in certain lysimeters or to seasonal changes.

For short-term time series analysis (seasonal cycles over 1-2 years), you might need to understand each variation and therefore it is useful to know exactly the composition of the pooled sample.

For long-term analysis (trend over 10 years), I assume that the pooled sample at a given depth is representative for the plot and that the temporal trend does not depend on the fact that one lysimeter had water or not in a given sampling period.

See you soon

Elisabeth

Permalink Reply by Tiina Maileena Nieminen on April 19, 2015 at 21:38

Dear all,

Thanks for your excellent  suggestions. I think you have created a very good basis for the developments needed.

 

See you soon in Göttingen!

 

Tiina