Database schema
as of 16 Feb 2015.
Data description tables
Stations
Example:
stationID | dcmipoint | name |
|
|---|---|---|---|
2 | name=Hyytiälä; east=24.294795; north=61.847463; elevation=179 | Hyytiälä |
|
Tags
Tags are used to link variables/columns to vocabularies or other standards.
Example:
tagID | vocabulary | tag | displaykeyword |
|---|---|---|---|
1 | air_temperature | ilma - lämpötila | |
2 | p19483 | ilma - lämpötila |
VariableTags are used to link tags to columns.
Events
Changes in data time series can be described with Events. They are linked to variables/columns via variableEvents that combine eventID and variableID.
Examples from some events linked to columns in table VAR_TREE:
eventID | event_type | event | period_start | period_end | who | etimestamp |
|---|---|---|---|---|---|---|
1 | ObjectChange | 2001-04-23 chambers installed on new pine shoots | 2001-04-23 00:00:00 | 2001-10-05 00:00:00 | Pertti Hari | 2014-11-24 10:17:49
|
19 | InstrumentConfiguration
| URAS 3 CO2 and H2O analyzers changed to Li-Cor Li-840 | 2010-04-16 00:00:00 | NULL | Toivo Pohja, | 2014-11-14 08:54:34 |
14 | NewVersionRelease | Post-calibration correction to 1999-2008 H2O concentration and transpiration data using tower RH vs recently calibrated chamber H2O analyser regression | 1999-05-07 00:00:00 | 2009-01-01 00:00:00 | Tapani Hopkins, | 2014-06-16 12:00:00 |
Fields period_start and period_end indicate the time frame when the event affects the data. Period_end may be left empty which means that the event is still "on". New event may or may not override older event of the same type even if period_end of the previous event is not determined, some responsibility for event interpretation is left to the data user...
Field etimestamp indicates when the event was inserted to the database or (in case of NewVersionRelease) when the data was updated.
It is possible to indicate the data life cycle events relevant to the database with very small number of event types:
1. Changes in data collection
a) change in the object being measured without setting up new column in the database, for instance, moving a sensor to another location in the forest (in many atmospheric/meteorological variables small horizontal movements are irrelevant and often not reported at all) or moving shoot chamber x to another pine shoot
b) sensor/instrument change or other significant change in the measurement configuration
c) calibration
2. Changes in data processing
a) change in data processing from period_start on while keeping older data as is, for instance, soil temperature until 2004 was calculated as average of 4 sensors, since 2005 average of 5 sensors
b) reprocessing of data, for instance, applying different quality check.
Corresponding event types:
1a) ObjectChange
1b) InstrumentConfiguration
1c) Calibration
2a) DataProcessing
2b) NewVersionRelease
The vocabulary is based on DDI data lifecycle events and ENVRI reference model.