Modelling the time period of a time series

Created by: JoachimBenner

Actually, we are using the ISO type TM_Period for beginning and end of both, regular and irregular time series, In the irregular time series, each value has an absolute time position (TP_Position), which means that the definition of a TemporalExtent in principle is superflous. The only use case could be to enable the validation, whether a certain value falls into the specified temporal extent. Is this really needed?

I propose to shift the the attribute _temporalExtent * from *TimeSeries to RegularTimeSeries.

In GML, TM_Period is encoded as gml:TimePeriodType. This type works very well to define absulute time periods, e.g.

1. January 2014 - 14. march 2014 2010 - 2012 January 2010 - march 2012 8.00 h - 12:00 h

What is not possible with gml:TimePeriodType is to specify a date interval without an explicit year, e.g.

January - October

1. January - 31. January

I think this is a problem, because e.g. statistical climate data do not refer to a specific year, but to specific days, years and months within a year. Actually, I do not know a good way to express this type of information in GML. Has anyone ideas?

I would like to label this issue, but I do not know how to do this!